# Tolz.me > Free browser-based tools for Discord, text, images, developer tasks, social media, and gaming. No account required. This file holds the full text of every tool page and every guide, written for AI agents and retrieval pipelines. ## Site summary Tolz.me is a free tools site. Every tool lives on its own page and does one job: build a Discord timestamp, format a message, convert a cron expression, work out a chmod value, sort a list of lines, total shift hours, or pull a colour palette out of a photo. There is no account to create and nothing to install. Categories cover Discord, text, image, developer, social media, calculator, and gaming. Each tool page carries the same structure: a short description, an explanation of how the tool works, FAQs, use cases, and related links. Longer guides live in the blog and go deeper than the tool pages do. ## MCP server - Server URL: https://tolz.me/api/mcp - Transport: Streamable HTTP - Authentication: none - Access: read-only. Every tool is deterministic, so the same input always returns the same output, and no tool changes anything on the server. - Tools: 52 site tools, plus two catalog tools. ### list_tools Returns every tool on the site with its name, category, one-line description, and page URL. Pass an optional category to narrow the list, for example `category: "discord"`. ### search_blog Searches the guides by title, description, and tag. Pass a query, for example `query: "cron"`. ### Examples - "List the Tolz.me tools in the developer category and give me the URLs." - "Search the Tolz.me guides for cron expressions." - "Convert the cron expression 0 9 * * 1-5 into plain English." - "Work out the chmod value for a script that the owner can run and everyone else can read." ## Tools ### Discord Timestamp Generator - URL: https://tolz.me/discord/timestamp-generator - Category: discord - Description: Create Discord timestamps that display in each user's local timezone with seven format styles. Pick a date and time, choose a display style, and copy a Discord-ready timestamp tag. Discord renders the time in every reader's local timezone automatically. Seven styles available from short time to relative countdown. #### FAQs **Q: How do I use a Discord timestamp?** **A:** Copy the generated code and paste it into any Discord message. Discord renders it as a clickable date that adjusts to the reader's timezone. **Q: What timestamp formats does Discord support?** **A:** Seven styles: short time (t), long time (T), short date (d), long date (D), combined date-time (f and F), and relative time (R) for countdowns. **Q: Do Discord timestamps work on mobile?** **A:** Yes. Timestamps render on Discord desktop, mobile, and web apps when you paste the code into a message or embed description. **Q: What is the difference between f and F timestamp formats?** **A:** f shows the full date and time in a standard format like 'Saturday, January 6, 2024 3:30 PM'. F adds the day of the week at the start. **Q: Can I use timestamps in Discord embeds?** **A:** Yes. Paste the timestamp code into embed descriptions and field values. Embed titles do not support timestamps—use the description field. **Q: Why should I avoid reusing old timestamps for recurring events?** **A:** A Unix timestamp always points to one specific instant. Reusing an old timestamp for a new event makes it show the original date instead. **Q: Is this Discord timestamp generator free?** **A:** Yes. The tool runs entirely in your browser with no account or payment required. ### Discord Text Formatting Generator - URL: https://tolz.me/discord/text-formatter - Category: discord - Description: Apply bold, italic, underline, strikethrough, spoiler, and code formatting to Discord messages. Pick a formatting style, type your text, and copy the Discord markdown. Seven styles available: bold, italic, underline, strikethrough, spoiler tags, inline code, and code blocks. Preview updates as you type. #### FAQs **Q: What Discord formatting styles are available?** **A:** Seven styles: bold (**text**), italic (*text*), underline (__text__), strikethrough (~~text~~), spoiler (||text||), inline code (`text`), and code blocks (```text```). **Q: Does formatting work in Discord embeds?** **A:** Yes for descriptions and field values. Embed titles do not render markdown. Forum posts support all formatting styles. **Q: Can I combine multiple styles in one message?** **A:** Yes. You can stack styles like ***bold italic*** or ~~**strikethrough bold**~~. Test combinations on mobile before using them in important announcements. **Q: Do spoilers hide text on mobile?** **A:** Spoilers render as gray bars on both desktop and mobile. Readers tap or click to reveal the hidden text. **Q: How do headers work in Discord messages?** **A:** Start a line with # for a large heading, ## for medium, or ### for small. Headers work in messages and forum posts. **Q: What is the difference between inline code and code blocks?** **A:** Inline code wraps short snippets in single backticks. Code blocks use triple backticks on their own lines and preserve whitespace for multi-line content. **Q: Is this tool free?** **A:** Yes. Formatting runs in your browser with no login required. ### Discord Color Text Generator - URL: https://tolz.me/discord/color-text-generator - Category: discord - Description: Create colored Discord text using ANSI code blocks with 8 terminal colors. Pick a color from eight ANSI terminal options, enter your text, and copy the generated code block. Desktop Discord renders the color; mobile shows plain monospace text. Works in messages, forum posts, and embed descriptions. #### FAQs **Q: Do ANSI colors work on Discord mobile?** **A:** ANSI colors display on Discord desktop. Mobile clients show plain monospace text instead, so do not rely on color alone to convey meaning. **Q: How do I paste colored text in Discord?** **A:** Copy the full code block including the opening ```ansi and closing ```. Paste it directly into a Discord message. Removing the fences breaks the color rendering. **Q: Which colors are supported?** **A:** Eight standard ANSI colors: red, green, yellow, blue, magenta, cyan, white, and gray. **Q: Can I use multiple colors in one message?** **A:** Generate separate blocks for each color and post them as consecutive lines. Discord does not support multiple colors within a single ANSI code block. **Q: Can I combine ANSI colors with Discord markdown?** **A:** Not in the same line. ANSI blocks and Discord markdown operate independently. Use separate lines for colored text and bold or italic formatting. **Q: Where do ANSI blocks work?** **A:** Messages, forum posts, and embed descriptions. They do not work in embed titles or anywhere Discord strips code block formatting. **Q: Is this tool free?** **A:** Yes. The tool runs in your browser with no login required. ### Discord Font Generator - URL: https://tolz.me/discord/font-generator - Category: discord - Description: Transform text into Unicode fancy fonts for Discord display names, bios, and messages. Choose from five Unicode font styles—Bold Sans, Italic Sans, Monospace, Script, and Double Struck—and transform any text into stylized characters that Discord accepts anywhere normal text works. #### FAQs **Q: Do fancy fonts work in Discord usernames?** **A:** Discord usernames only allow lowercase a-z, 0-9, underscores, and periods. Fancy text works in display names, server nicknames, bios, and messages. **Q: What font styles are available?** **A:** Five styles: Bold Sans, Italic Sans, Monospace, Script, and Double Struck. Each maps letters to Unicode characters from a different alphabet block. **Q: Will everyone see my fancy font correctly?** **A:** Most modern devices render Unicode fonts properly. Some older devices or custom ROMs may show boxes for rare characters. **Q: Can I use fancy fonts in Discord bios?** **A:** Yes. Bios, display names, server nicknames, channel topics, and messages all accept Unicode text. Test on mobile if your audience uses varied devices. **Q: Is there a character limit?** **A:** Display names and nicknames max out at 32 characters. Very long fancy strings may truncate or be rejected by Discord's name filter. **Q: What happens to characters that cannot be transformed?** **A:** Letters without a mapping in the selected style pass through unchanged. Numbers, punctuation, and emoji also pass through as-is. **Q: Is this tool free?** **A:** Yes. The tool runs in your browser with no login required. ### Discord ASCII Art Generator - URL: https://tolz.me/discord/ascii-art-generator - Category: discord - Description: Generate block-letter ASCII art banners and random kaomoji emoticons for Discord. Convert short text into five-line block letters for banners, or pick a random kaomoji emoticon for chat reactions. Block art needs a code block in Discord to preserve alignment; kaomoji paste anywhere as plain text. #### FAQs **Q: What characters work in block letter mode?** **A:** Block letters support A-Z and 0-9. Spaces and punctuation are ignored. Maximum input is 12 characters. **Q: Should I wrap ASCII art in a code block?** **A:** Yes, for block art. Wrap in triple backticks so Discord uses a monospace font and preserves column alignment. **Q: What are kaomoji?** **A:** Kaomoji are Japanese-style text emoticons like (╯°□°)╯ and ʕ•ᴥ•ʔ. They paste as plain text and work anywhere emoji do. **Q: Will ASCII art work on mobile Discord?** **A:** Wide art may wrap on mobile. Keep input short (5-7 characters) for best mobile results. **Q: Can I use block art in announcements?** **A:** Block art works in any text field. Use it for event banners, tournament headers, or decorative separators in pinned messages. **Q: How do I get a different kaomoji?** **A:** Click Randomize kaomoji to cycle through options. Each click picks a new emoticon from the curated library. **Q: Is this tool free?** **A:** Yes. The tool runs in your browser with no login required. ### Discord Username Generator - URL: https://tolz.me/discord/username-generator - Category: discord - Description: Generate aesthetic, funny, and minimalist Discord username ideas. Pick a style, click Generate, and copy a Discord-ready username. Three word lists cover aesthetic, funny, and minimalist vibes. Availability is checked in your Discord account settings after you copy. #### FAQs **Q: Does this check if a username is available on Discord?** **A:** No. Copy a name you like and try it in User Settings, Profiles. Discord tells you if it is taken. Have a few backups ready before you open settings. **Q: What username styles are available?** **A:** Aesthetic combines soft words and underscores like misty_dreams42. Funny stacks titles and nouns like CaptainPotato. Minimalist keeps it short like neolux0 or zen_ryu. **Q: Can I use these names as server nicknames?** **A:** Yes. Generated strings work as your global username or per-server display nickname. The username rules apply to the global handle; display names are more flexible. **Q: What characters are valid in a Discord username?** **A:** Only lowercase a-z, digits 0-9, underscores _, and periods . are allowed. The period cannot be first or last, and no two periods can be consecutive. 2-32 characters. **Q: Are generated names unique?** **A:** Names are random combinations from word lists. Uniqueness is not guaranteed. The period rule and 32-character max apply to all outputs. **Q: Can I edit the generated name before copying?** **A:** The output is a single text string. Edit it before copying as long as you stay within Discord's character rules. Adding suffix numbers is the most common tweak. **Q: Is this tool free?** **A:** Yes. Generation runs in your browser with no account needed. ### Discord Bio Template Generator - URL: https://tolz.me/discord/bio-template - Category: discord - Description: Copy aesthetic Discord bio templates with one click. Choose from five bio templates, edit the text to match your interests, and paste the result into your Discord profile About Me field. Five styles cover gaming, aesthetic, creator, student, and minimal layouts. #### FAQs **Q: Where do I paste my Discord bio?** **A:** Open User Settings, Profiles, click the pencil icon on About Me, and paste. Save to apply. Check your profile card on mobile to confirm it displays correctly. **Q: Can I edit templates before copying?** **A:** Yes. The textarea is fully editable. Change hobbies, replace placeholder links with your actual handles, and remove emoji you do not want. **Q: Do Discord bios support markdown formatting?** **A:** Plain text with line breaks is the most reliable. Masked links like [text](https://url.com) render on desktop but not always on mobile. Bold, italic, and headers do not render in bios. **Q: How long can a Discord bio be?** **A:** Discord does not publish a hard limit. Around 190 characters shows fully on mobile before truncation. Keep bios concise for full visibility. **Q: Can I use Unicode fonts in my bio?** **A:** Yes. Unicode text from the font generator works in bios, display names, and server nicknames. Test on mobile to confirm characters render correctly. **Q: Do server profiles override my bio?** **A:** Some servers set custom per-server profiles that override the global bio in that context. Your main bio stays in place elsewhere. **Q: Is this tool free?** **A:** Yes. Templates generate in your browser with no account required. ### Discord Role Name Generator - URL: https://tolz.me/discord/role-name-generator - Category: discord - Description: Generate creative Discord role names for staff, VIP, and color roles. Pick a role tier, click Generate, and copy a name ready for Server Settings. Five themes: Admin, Moderator, Member, VIP, and Color roles. Names are cosmetic — set permissions separately in role settings. #### FAQs **Q: How do I create a role with the generated name?** **A:** Go to Server Settings, Roles, click Create Role, paste the name, pick a color, and set permissions. Save when done. **Q: Do role names grant permissions?** **A:** No. Names are labels only. Permissions are set with checkboxes in role settings. Use the permission calculator to get exact integer values for bot configurations. **Q: What themes are available?** **A:** Admin tier: Head Admin, Server Owner, Director, Overseer. Moderator: Moderator, Guardian, Enforcer, Peacekeeper. Member: Member, Regular, Citizen, Resident. VIP: VIP, Elite, Premium, Patron. Color: Crimson, Azure, Emerald, Violet, Gold. **Q: Can I rename an existing role?** **A:** Yes. Click the role in Server Settings, Roles, and change the name. It updates for all members with that role immediately. **Q: What is the character limit for role names?** **A:** Discord allows up to 100 characters in role names. For the member list on mobile, shorter is better — 15 characters or fewer avoids truncation. **Q: Why do color roles need distinctive names?** **A:** Members see the role name in the sidebar. Azure Crimson reads clearly. Abstract names like Twilight require members to check the actual color swatch to understand what they picked. **Q: Is this tool free?** **A:** Yes. Generation runs in your browser with no login required. ### Discord Dividers Maker - URL: https://tolz.me/discord/dividers-maker - Category: discord - Description: Copy stylish divider lines for Discord bios and messages. Pick a divider style and copy a Unicode separator line for Discord bios, rules posts, pinned announcements, and forum content. Eight styles from simple lines to decorative sparkles. Paste on its own line between text sections. #### FAQs **Q: Where can I use dividers on Discord?** **A:** Bios, rules channels, pinned messages, forum post content, and embed descriptions. They render as Unicode text on desktop and mobile. **Q: Will long dividers break on mobile?** **A:** Narrow screens may wrap very long rows. Shorter styles like Simple line, Dotted, and Hearts handle mobile width better. Test on your device if the visual matters. **Q: Do dividers work in channel topics?** **A:** Yes but very long rows can get truncated on mobile. Keep dividers in topic lines short or use a simpler style for topics. **Q: Can I combine dividers with markdown?** **A:** Dividers are plain Unicode characters. They do not conflict with markdown. Place them on their own line between sections for the cleanest result. **Q: Do screen readers handle dividers well?** **A:** Screen readers announce Unicode symbols character by character. Pair dividers with clear section labels rather than relying on them for structure. **Q: How do I choose the right style?** **A:** Simple line works everywhere. Stars and Hearts fit aesthetic or community servers. Sparkle is decorative for creative content. Block stands out for major section breaks. **Q: Is this tool free?** **A:** Yes. Copy as many divider styles as you need with no login. ### Discord Emoji Maker - URL: https://tolz.me/discord/emoji-maker - Category: discord - Description: Resize images to 128×128 or 256×256 for Discord custom emojis. Upload any image, pick a size, and download a square PNG ready for Discord upload. Center-crop keeps your image focused. 128×128 for standard emoji, 256×256 for animated slots on boosted servers. All processing happens in your browser. #### FAQs **Q: What size does Discord require for custom emoji?** **A:** Static emoji upload at 128×128 pixels. Animated emoji on boosted servers use 256×256. Both must be under 256 KB file size. **Q: Is my image uploaded to your servers?** **A:** No. All resizing runs in your browser using the Canvas API. Your image never leaves your device. **Q: How does center-crop work?** **A:** The tool scales your image to cover the square canvas, then crops the edges to reach the target size. The center of your image is preserved. Keep important content centered in your source art. **Q: What file types can I upload?** **A:** Any format your browser can render as an image: PNG, JPG, GIF, WebP, BMP, and others. GIF uploads work for static emoji — use a separate GIF editor for animation. **Q: Why does my exported emoji look cropped?** **A:** Center-crop removes edges. If your main subject was off-center in the source, it may have been cropped out. Use a square source image or reposition the important content in the center before uploading. **Q: Can I create animated emoji with this tool?** **A:** This tool outputs a static PNG. To make animated emoji, create a GIF separately and resize it using the 256×256 setting before uploading to Discord. **Q: Is this tool free?** **A:** Yes. Resize and download without an account or payment. ### Discord Emoji Resizer: How Do I Make an Emoji 128x128 Under 256 KB? - URL: https://tolz.me/discord/emoji-resizer - Category: discord - Description: Resize a PNG, JPG, WebP or GIF to 128×128 and keep it under Discord's 256 KB emoji limit. Animated GIFs included. The Discord emoji resizer turns any PNG, JPG, WebP or GIF into a 128 by 128 file that sits under Discord's 256 KB upload limit. Still images export as PNG with transparency kept, and animated GIFs are decoded frame by frame, resized, and re-encoded with their original delays. A size bar shows exactly where your file lands against the limit, and the tool shrinks the palette or drops frames when it has to. Everything runs in your browser. #### FAQs **Q: What size does Discord need for a custom emoji?** **A:** 128 by 128 pixels. Discord keeps the emoji as a 128x128 image and scales anything larger, which softens the art. Uploading the exact square avoids that step. **Q: How do I get an emoji file under 256 KB?** **A:** This tool measures the output against Discord's 256 KB limit and shrinks it until it fits. It reduces the color palette first, then drops frames from an animation and stretches the delays so the timing stays the same. **Q: Is my image uploaded to your servers?** **A:** No. Decoding, resizing, and re-encoding all run in your browser using the Canvas API. Your file never leaves your device. **Q: Does it handle animated GIFs?** **A:** Yes. Every frame is decoded, composited with its disposal rules, resized to 128 by 128, and re-encoded with the original frame delays. **Q: What is the difference between contain and cover?** **A:** Contain fits the whole image inside the square and pads the empty space with transparency. Cover fills the square and crops the edges that do not fit. **Q: Why does my small image look blurry?** **A:** Upscaling cannot invent detail. A 32x32 icon stretched to 128x128 is smoothed by the browser and looks soft. Start from art that is at least 128 by 128. **Q: Can I upload a WebP or a JPG?** **A:** Yes. Discord accepts JPEG, PNG, GIF, WebP, and AVIF for emoji. This tool exports PNG for still images and GIF for animations, and offers a WebP download for still images. ### Discord Sticker Compressor: How Do I Get a Sticker Under 512 KB at 320x320? - URL: https://tolz.me/discord/sticker-compressor - Category: discord - Description: Compress a PNG, APNG, GIF, JPG or WebP into a 320×320 sticker that fits Discord's 512 KB limit. Animated stickers included. The Discord sticker compressor turns any PNG, APNG, GIF, JPG or WebP into a 320 by 320 file that sits under Discord's 512 KB sticker limit. Still images come out as PNG, and animations are decoded frame by frame and re-encoded as APNG, then trimmed to five seconds because that is the cap Discord puts on animated stickers. A size bar shows where your file lands against the limit, and the tool shrinks the palette or drops frames when it has to. Everything runs in your browser. #### FAQs **Q: What size does Discord need for a sticker?** **A:** 320 by 320 pixels. Discord scales anything else, and a scaled sticker loses the sharp edges of your art. **Q: How do I get a sticker file under 512 KB?** **A:** The tool measures the output against Discord's 512 KB limit and shrinks it until it fits. It reduces the color palette first, then drops frames from an animation and stretches the delays so the timing stays the same. **Q: Is my file uploaded to your servers?** **A:** No. Decoding, resizing and encoding all run in your browser using the Canvas API. Your file never leaves your device. **Q: What happens to an animated GIF?** **A:** Every frame is decoded, composited with its disposal rules, resized to 320 by 320 and re-encoded as an APNG, the animated PNG format Discord accepts. **Q: How long can an animated sticker be?** **A:** Five seconds. The trim switch cuts a longer animation to the first five seconds. Turn the switch off and you get the full animation with a warning that Discord will reject it. **Q: Can this make a Lottie sticker?** **A:** No. Lottie is a JSON animation file and Discord only accepts it on verified or partnered servers. This tool makes PNG and APNG, which work on every server. **Q: What do I have to type on the upload screen?** **A:** The name needs 2 to 30 characters. The description can be left empty or filled with 2 to 100 characters. The tags can hold up to 200 characters. **Q: Why does my sticker look blurry?** **A:** Upscaling cannot invent detail. A 64x64 icon stretched to 320x320 is smoothed by the browser and looks soft. Start from art that is at least 320 by 320. ### Discord Banner Resizer: What Size Is a Discord Banner? - URL: https://tolz.me/discord/banner-resizer - Category: discord - Description: Resize a PNG, JPG, WebP or GIF to a Discord banner size: 960×540 server banner, 680×240 profile banner, 600×240 per-server profile, or 1920×1080. The Discord banner resizer fits a PNG, JPG, WebP or GIF to a banner size Discord actually uses: 960 by 540 for a server banner, 1920 by 1080 for the high resolution version, 680 by 240 for your own profile banner, and 600 by 240 for a per-server profile banner. Cover mode crops the picture and two focus sliders decide what stays, while contain mode keeps the whole picture and pads the empty space with a background colour. Still images come out as PNG or JPG, animated GIFs stay animated GIFs, and the tool warns you when the file goes past the 10 MB cap or when the source is too small to fill the frame. Everything runs in your browser. #### FAQs **Q: What size is a Discord server banner?** **A:** 960 by 540 pixels in a 16:9 shape. Discord also accepts 1920 by 1080 and scales that down for you. The server needs Boost Level 2 before a banner option appears at all. **Q: What size is a Discord profile banner?** **A:** At least 680 by 240 pixels. That is the minimum in Discord's Custom Profiles support article. A lot of older sites still quote 600 by 240, which is the size of the per-server profile banner. **Q: Do I need Nitro for a profile banner?** **A:** Yes, if you want an image or GIF banner on your own profile. Discord's Custom Profiles article lists Nitro as the requirement for that. The default colour themes work without it. **Q: What is a per-server profile banner?** **A:** A second profile banner that only shows up inside one server, so you can use different artwork in different communities. It is 600 by 240 pixels, which is where the older 600 by 240 figure comes from. **Q: Can I use an animated Discord banner?** **A:** Yes. Drop in an animated GIF and the download stays an animated GIF with the original frame delays. Discord plays an animated server banner for about five seconds and then pauses it until the pointer moves over it. **Q: What is the file size limit for a Discord banner?** **A:** Profile banner uploads are capped at 10 MB. This tool measures the finished file and warns you when it goes over that line, so you can switch to JPG or shrink the source. **Q: Does the tool draw the areas Discord covers?** **A:** No. The preview marks the server title strip and the avatar overlap in the lower-left corner so you can check them, and those marks are never part of the download. **Q: Is my image uploaded to your servers?** **A:** No. Decoding, resizing and encoding all run in your browser using the Canvas API. Your file never leaves your device. ### Discord Server Name Generator - URL: https://tolz.me/discord/server-name-generator - Category: discord - Description: Generate themed Discord server name ideas for gaming, community, study, and creative groups. Pick a theme — gaming, community, study, or creative — and generate server name ideas from curated word lists. Click Generate as many times as you need until something fits. The tool does not check availability; Discord allows duplicate display names. Set the name in Server Settings, Overview. #### FAQs **Q: Can I change my server name after creating it?** **A:** Yes. Edit the name anytime in Server Settings, Overview. The change propagates to existing invite links within seconds. **Q: What themes are available?** **A:** Four themes: Gaming, Community, Study, and Creative. Each draws from its own word list tailored to that server type. **Q: Does this check if a name is taken?** **A:** No. Discord server names do not need to be unique across the platform. Pick any name that fits your community. A vanity URL (discord.gg/yourname) does need to be available and requires a Server Boost. **Q: How long can a Discord server name be?** **A:** Discord allows up to 100 characters, but names over 30-40 characters get truncated on mobile invite previews. Shorter is usually better. **Q: Can I combine parts of different generated names?** **A:** Yes. Generate until you find parts you like, then manually combine them. The generator gives starting points, not fixed outputs. **Q: Do server names support special characters?** **A:** Spaces, hyphens, and standard Latin characters work. Discord may strip or reject characters outside that set. Stick to basic text for maximum compatibility. **Q: Is this tool free?** **A:** Yes. Generate as many name ideas as you want with no account required. ### Discord Server Template - URL: https://tolz.me/discord/server-template - Category: discord - Description: Export channel and category layouts for gaming, community, and study Discord servers. Choose a gaming, community, or study layout and export a category-and-channel checklist as markdown or JSON. Work through the list in Server Settings to build your server systematically. The markdown format is a readable planning document; the JSON format works for bot configuration or documentation. #### FAQs **Q: Does this create channels in my Discord server automatically?** **A:** No. This is a planning and reference tool. You create each category and channel manually in Server Settings. The output is a checklist to work from. **Q: What layouts are available?** **A:** Three templates: Gaming (with LFG and voice squad channels), Community (with introductions and casual chat), and Study (with focus rooms and homework help). **Q: What export formats are available?** **A:** Checklist (markdown) gives you a readable outline with category headers and channel bullets. JSON gives you a raw data structure for bot configs or version-controlled documentation. **Q: Can I edit the channel list before building?** **A:** Yes. Copy the output and remove, rename, or reorder channels before you start creating them in Discord. Empty channels look abandoned, so only create what you will use. **Q: Should I set permissions before or after creating channels?** **A:** Before. Set category permissions first since they cascade to all channels inside the category. Then adjust individual channel permissions as needed. **Q: Is this tool free?** **A:** Yes. Export as many layouts as you need with no account required. ### Discord Server Rules Template - URL: https://tolz.me/discord/server-rules-template - Category: discord - Description: Copy and customize Discord server rules for basic, gaming, or community servers. Pick a rule template — Basic, Gaming, or Community — and edit the text in the textarea before copying. Paste the result into a read-only #rules channel and pin it. Templates cover respect, spam, channel use, and common issue areas; customize every line to match your actual server culture. #### FAQs **Q: Are these rules legally binding?** **A:** No. Templates are starting points for community guidelines, not legal documents. Edit every line to reflect your actual server rules and local laws. **Q: Where should I post server rules?** **A:** Create a #rules channel in your INFO or WELCOME category. Set it read-only for members, paste the rules, and pin the message. Link to it from your welcome message using a channel mention. **Q: Can I add or remove rules from the template?** **A:** Yes. The textarea is fully editable. Add rules for issues specific to your community, remove ones that do not apply, and reorder points as needed. **Q: Which template should I start with?** **A:** Basic suits general social servers. Gaming adds anti-cheat, LFG etiquette, and spoiler policies. Community focuses on debate boundaries, privacy, and reporting expectations. **Q: Do Discord rules posts support markdown formatting?** **A:** Yes. Use **bold** for rule titles, numbered lists for each point, and ## headers to separate sections. Masked links like [Report Form](https://link) look cleaner than raw URLs. **Q: How do I handle rule violations?** **A:** Match enforcement to server size. Small servers may rely on direct conversation. Larger servers should document a strike system separately so members know the consequences of repeated violations. **Q: Is this tool free?** **A:** Yes. Copy and edit as many rule sets as you need with no login required. ### Discord Welcome Message Template - URL: https://tolz.me/discord/welcome-message-template - Category: discord - Description: Copy and customize Discord welcome messages for new member onboarding. Choose a welcome message template, edit the text to match your channels and tone, and paste it into your welcome bot configuration or welcome channel. The editable textarea lets you replace {user} with your bot's mention syntax, update channel names, and adjust the greeting before copying. #### FAQs **Q: Does this send welcome messages automatically?** **A:** No. This tool generates the message text only. Copy the result and paste it into your welcome bot settings, or post it manually in a welcome channel. **Q: What does {user} mean in the template?** **A:** It is a placeholder for the new member's name or mention. In bot configurations, replace it with your bot's mention variable such as {user.mention} or {mention}. For manual posting, replace it with a literal greeting. **Q: Which template is best for large servers?** **A:** The Minimal template is short and works well for servers with automated onboarding. The Structured template uses a bullet-point list for start-here steps, which suits servers with role assignment or verification gates. **Q: Can I add a Discord timestamp to a welcome message?** **A:** Yes. Use the timer message generator or timestamp generator tool to create a timestamp tag, then paste it into the welcome message template before copying. **Q: Can I edit the channel names in the template?** **A:** Yes. The textarea is fully editable. Replace #rules and #roles with your actual channel names. If you do not have a roles channel, remove that line. **Q: Do welcome messages support Discord markdown?** **A:** Yes. Bold, bullet points, and channel mentions all render in welcome messages. Use **bold** on the server name or key instructions for scannability. **Q: Is this tool free?** **A:** Yes. Generate and copy welcome messages with no account required. ### Discord Timer Message Generator - URL: https://tolz.me/discord/timer-message-generator - Category: discord - Description: Build Discord announcement messages with embedded local-time timestamps for events, giveaways, and meetings. Pick a template, set a date and time, and copy a message with a Discord timestamp already inserted. Discord renders the timestamp in each reader's local timezone automatically. This tool generates text only — use a bot or Discord's built-in scheduled events to handle actual delayed delivery. #### FAQs **Q: Does this schedule or send messages automatically?** **A:** No. This generates the message text with a Discord timestamp tag embedded. To schedule delivery, use a bot like Carl-bot or Discord's built-in scheduled events feature. **Q: How does the timestamp work?** **A:** The tool creates a Discord `` tag from your date and time input. When pasted into Discord, it renders as a human-readable date adjusted to each member's local timezone. **Q: What display style does the timestamp use?** **A:** Style F by default — the full date and time including the day of the week, like Saturday, January 6, 2024 3:30 PM. You can change the style letter manually after copying if you prefer short time (t), relative countdown (R), or another format. **Q: What templates are available?** **A:** Three: Event reminder (with a voice channel reference), Giveaway (with an end time), and Meeting (short and direct). All are editable in the textarea before copying. **Q: Why should I not reuse old timestamps for recurring events?** **A:** A Unix timestamp always points to one specific instant. If you copy a timestamp for March 15 and reuse it for April's event, the rendered date still shows March 15. Generate a fresh timestamp each time you announce a recurring event. **Q: What timezone does the datetime picker use?** **A:** Your browser's local timezone. If your event is in a specific timezone different from yours, calculate the UTC equivalent and enter that as your local time. **Q: Is this tool free?** **A:** Yes. Generate announcement text with no account or payment required. ### Discord Permissions Calculator: What Is My Permission Integer? - URL: https://tolz.me/discord/permission-calculator - Category: discord - Description: Tick Discord permissions to build the role integer, or paste an integer to update the boxes. Both ways work live, and the result turns into a bot invite link. This Discord permission calculator works in both directions. Tick the permissions a role needs and it adds up Discord's official bit table into one integer, or paste an integer in decimal or hex and the checkboxes rebuild themselves from it. The list covers the 52 permissions Discord defines today, grouped the way Server Settings groups them. Add a client ID to turn the result into a bot install link, and see any bits Discord does not recognise before you save the value to a config file. #### FAQs **Q: What is a Discord permission integer?** **A:** One number that holds every permission a role has. Each permission owns one bit in it, so a role that can send messages and kick members carries two bits set. Discord stores role permissions this way, and bots read the same number to know what they may do. **Q: How do I use the permission calculator?** **A:** Both fields are live. Tick a box and the integer field rewrites itself with the new total. Type or paste an integer in decimal or with a 0x hex prefix and the checkboxes update as you go. The output row shows the value in decimal and hex, plus a count of the permissions you have on. **Q: What does the Administrator permission do?** **A:** Administrator grants every permission and bypasses channel overwrites, so a role with that bit ignores deny rules on a private channel. Keep it for a small owner tier. A moderator preset without that bit covers most moderation work. **Q: Why does my value contain bits Discord does not recognise?** **A:** Bit 47 is unused in Discord's table, and Discord adds new permissions by taking the next free bit, so values from a newer API version can carry bits this table does not name. The tool lists those bits in a warning instead of dropping them silently. **Q: Should I store a permission value as a number or a string?** **A:** As a string. Large combinations pass 2^53, the point where normal JavaScript numbers lose precision, and the Discord API sends the permissions field as a string. Keep it quoted in config files. **Q: How do I build a bot invite link with these permissions?** **A:** Enter your application client ID, a snowflake of 17 to 20 digits from the Developer Portal. The tool builds https://discord.com/oauth2/authorize with your client ID, the current integer, and the bot plus applications.commands scopes. Copy it or open it straight from the tool. **Q: Are channel overwrites included in the value?** **A:** No. The integer holds role level permissions only. Channel specific allow and deny overwrites are stored on the channel and managed inside each channel's permission settings. **Q: Is the Discord permission calculator free?** **A:** Yes. The arithmetic runs in your browser with BigInt, nothing is uploaded, and no account is needed. ### Discord ID Lookup - URL: https://tolz.me/discord/id-lookup - Category: discord - Description: Decode Discord snowflake IDs to find creation dates, and fetch public user or server details when Discord grants access. Paste any Discord snowflake ID and the tool decodes the creation timestamp instantly in your browser. When our bot has access to the entity, it also fetches the current username, avatar, and server name. Enable Developer Mode in Discord's settings to get Copy ID in right-click menus. #### FAQs **Q: How does Discord snowflake decode work?** **A:** Discord IDs encode a millisecond-precision Unix timestamp in the first 42 bits. The tool extracts those bits and converts them to a human-readable date. This calculation runs in your browser — no API call is needed for the decode. **Q: Why does the username or server name not appear?** **A:** Our bot needs to share a server with the user, or be a member of the server, for Discord to return live profile data. The snowflake decode always succeeds regardless of access. **Q: How do I get a Discord ID to look up?** **A:** Enable Developer Mode in User Settings, Advanced. Then right-click any username, message, channel, role, or server name and select Copy ID. Channel and role IDs require Developer Mode to appear in the context menu. **Q: Can I look up channel or role IDs?** **A:** The snowflake decode works for any valid Discord snowflake, including channels and roles. Live API details are limited to users and servers at this time. **Q: Is the ID lookup safe to use?** **A:** Yes. The snowflake decode is entirely client-side. API lookups use our bot's read-only access and do not store or log the IDs you search. **Q: Why does a very old account have a low ID number?** **A:** Discord IDs increment over time. Early accounts from 2015 have much smaller numbers than accounts created in 2024. The numeric value correlates with account age. **Q: Is Discord ID Lookup free?** **A:** Yes. Snowflake decoding runs instantly in your browser with no account required. Live lookups use our server-side bot. ### Discord Snowflake Decoder: When Was This ID Created? - URL: https://tolz.me/discord/snowflake-decoder - Category: discord - Description: Decode one or many Discord snowflake IDs to get the exact creation date, Unix timestamps, worker details, and bit breakdown. A Discord snowflake decoder reads the millisecond timestamp hidden in the top bits of any Discord ID and shows the exact creation date for it. Paste one ID or a whole list: each one comes back with your local time, UTC, Unix seconds and milliseconds, a paste-ready Discord timestamp, and the internal worker, process, and increment values behind the ID. #### FAQs **Q: What does a Discord snowflake decoder read?** **A:** It reads the 64-bit layout Discord uses for every ID. The top 42 bits hold a millisecond timestamp counted from the Discord epoch of 1420070400000, which is 1 January 2015. The remaining bits hold an internal worker ID, a process ID, and an increment counter. **Q: How do I get a Discord ID to decode?** **A:** Enable Developer Mode in Discord under User Settings, Advanced. Then right-click a username, message, channel, role, or server name and choose Copy ID. Channel and role IDs only show up in that menu once Developer Mode is on. **Q: Can I decode several IDs at once?** **A:** Yes. Paste one ID per line and every line gets its own result card. Blank lines are skipped, and the tool reports how many IDs decoded and how many need attention. **Q: Why does my ID show a date in the future?** **A:** A real Discord object cannot be created in the future, so the decoder rejects that value. It usually means a typo, or an extra digit added to the end of a valid ID, which multiplies the number by ten and pushes the timestamp far ahead. **Q: What are the worker ID, process ID, and increment?** **A:** They are the last 22 bits of the ID and they track Discord's internal machinery. The worker and process IDs point to the process that issued the ID, and the increment counts how many IDs that worker handed out. Two IDs sharing the same worker and process were issued close together. **Q: Why does Discord send IDs as strings?** **A:** A snowflake is a 64-bit number and can be larger than the biggest integer JavaScript stores exactly. If you parse one as a number you lose the trailing digits and land on a date that is off by months, so the API keeps IDs as strings and this decoder uses BigInt. **Q: Is the Discord Snowflake Decoder free?** **A:** Yes. Every decode runs in your browser, no account is needed, and the IDs you paste are not sent anywhere. ### Discord Profile Viewer - URL: https://tolz.me/discord/profile-viewer - Category: discord - Description: Preview your Discord profile by connecting your account, or look up public avatar and name data for any user ID. Connect your Discord account via OAuth to see your own avatar, banner, display name, username, pronouns, accent color, and bio as others see them. Look up any user ID to see their public avatar and name when our bot shares a server with them. Bios and pronouns for other users are not accessible through Discord's API. #### FAQs **Q: What does Connect Discord do?** **A:** It starts Discord's OAuth flow and requests only the identify scope — read-only access to your public profile. We do not post messages, access your friends, or modify your account in any way. Disconnect clears the session token immediately. **Q: Can I view another user's bio or pronouns?** **A:** No. Discord's API does not expose another account's bio, pronouns, or accent color to third-party applications, even when our bot shares a server with them. You can only see your own full profile this way. **Q: What does the ID lookup show?** **A:** For any valid user ID, the tool shows the account creation date (from the snowflake decode) and, when our bot has access, the avatar image, display name, and username. The bio is not available. **Q: Why does the lookup fail for some users?** **A:** Our bot must share at least one server with the user for Discord to return profile data. If the user has blocked the bot, is in a private server the bot cannot access, or has privacy settings that restrict API lookups, the result is an error after the snowflake decode. **Q: Is my Discord login stored on your servers?** **A:** No. We store a short-lived access token in a browser cookie to load your profile during the session. It is cleared when you disconnect or close the browser. We do not store your Discord credentials or profile data on our servers. **Q: How do I find a user ID to look up?** **A:** Enable Developer Mode in Discord's Advanced settings, then right-click any username or message and select Copy User ID. You can also use the ID Lookup tool to decode any snowflake you already have. **Q: Is the Discord Profile Viewer free?** **A:** Yes. Your own profile loads via OAuth with no payment. ID lookups run with our bot. ### Discord Channel Name Generator - URL: https://tolz.me/discord/channel-name-generator - Category: discord - Description: Generate themed channel names for general chat, gaming, voice, info, and moderation spaces. Pick a theme and click Generate. Pick a channel theme — general chat, gaming, voice, info, or moderation — and generate a random name from that theme's word list. Channel names must use lowercase letters, numbers, and hyphens. Discord converts spaces to hyphens when you create the channel. Copy the name and paste it into Server Settings, Channels. #### FAQs **Q: What characters are valid in a Discord channel name?** **A:** Lowercase letters a-z, digits 0-9, and hyphens. Spaces are not allowed — Discord turns them into hyphens automatically. Underscores are technically valid but not standard for channel names. **Q: How do I use the generated name in Discord?** **A:** Open Server Settings, Channels, click the plus icon to create a channel, paste the name into the Channel Name field, and save. Discord accepts hyphens as word separators. **Q: Can I use these names for categories?** **A:** Categories are separate from channels and allow spaces, Unicode characters, and longer text. You can name a category "Gaming" and put multiple gaming-themed channels inside it. **Q: How many channels should a new server start with?** **A:** A small server needs at minimum a general chat, a rules channel, and maybe an introductions or welcome channel. Add more as the community grows. Empty channels feel abandoned. **Q: Can I rename a channel later?** **A:** Yes. Go to Server Settings, Channels, hover over the channel, and click the pencil icon. Changing the name breaks any existing invite URLs that referenced the old name. **Q: What theme should I pick for a voice-only channel?** **A:** The Voice theme generates names like General Voice, Chill VC, Music Lounge, AFK Room, and Squad Voice. These work as labels that complement the member list inside the voice channel. **Q: Is the channel name generator free?** **A:** Yes. Generate and copy as many names as you need with no login required. ### Discord Mention Formatter - URL: https://tolz.me/discord/mention-formatter - Category: discord - Description: Format Discord mentions for users, roles, channels, and custom emoji from their numeric IDs. Copy the exact syntax for messages, bots, and webhooks. Pick a mention type — user, role, channel, or custom emoji — enter the numeric ID, and copy the formatted syntax. User mentions are <@id>, role mentions are <@&id>, channel mentions are <#id>, and emoji are <:name:id> or for animated. Enable Developer Mode in Discord's settings to get Copy ID in right-click menus. #### FAQs **Q: How does the mention formatter work?** **A:** Select a mention type, enter the numeric ID, and the tool builds the correct bracket syntax. Copy the result and paste it into any Discord message, thread, forum post, or embed description. **Q: What is the difference between user and role mention syntax?** **A:** User mentions use <@id> — the at-sign alone. Role mentions add an ampersand: <@&id>. The ampersand is the distinguishing character. Using <@> on a role ID renders as a user mention, not a role ping. **Q: Where do I get the ID for a custom emoji?** **A:** Right-click the emoji in the emoji picker or in a message and select Copy ID. You also need the emoji's name as it appears on that server. For animated emoji, check the Animated box in the tool. **Q: Why does my emoji mention show as plain text?** **A:** The emoji must be available on a server your bot or webhook is in. If your server does not have that emoji — either because it is from another server you do not share or because you lack Nitro — Discord renders the raw tag instead of the emoji image. **Q: Can I use mentions inside code blocks?** **A:** Yes. Discord processes mention syntax before applying code block formatting, so <#channel-id> still renders as a clickable channel link even inside a code block. Inline code spans work the same way. **Q: How do I find IDs to use with this tool?** **A:** Enable Developer Mode in User Settings, Advanced. Then right-click any user, role, channel, or emoji in Discord and choose Copy ID. The ID appears as a long number in the context menu. **Q: Is the Discord Mention Formatter free?** **A:** Yes. Formatting runs entirely in your browser with no login required. ### Discord Embed Color Picker - URL: https://tolz.me/discord/embed-color-picker - Category: discord - Description: Convert hex colors to Discord embed decimal color values. Pick a color visually and copy the decimal RGB value Discord embeds need, plus the matching hex code. The live preview shows how the color looks as an embed stripe. #### FAQs **Q: Why does Discord use decimal instead of hex for embed colors?** **A:** Discord's API accepts embed colors as a decimal integer. If your bot library or webhook expects a number, use the decimal. If it expects a string, use the hex code. Using the wrong format is the most common reason embed colors fail to render. **Q: What is Discord blurple in decimal?** **A:** #5865F2 converts to 5793266. That is the default when you open the color picker. **Q: Can I paste a hex code directly into the picker?** **A:** The color input is a visual picker, not a text field. To use a specific hex, pick a color close to it first, then note the hex output and fine-tune from there. For exact hex matching, open the picker in a design tool and use the same color. **Q: Do both the decimal and hex outputs update at the same time?** **A:** Yes. Move the picker and both values update instantly. Copy whichever format your destination requires. **Q: Can I use this to match my server's role colors?** **A:** Open your server icon in an image editor, use its color picker on the main color, and note the hex value. Then use this picker to convert it to the decimal Discord uses for embeds. **Q: Does the preview accurately show the embed stripe?** **A:** The preview uses a 6-pixel left border on a white card, which matches how Discord renders the stripe on desktop. Mobile clients may render slightly narrower stripes. The color itself is accurate. **Q: Is this tool free?** **A:** Yes. All conversions run in your browser with no account needed. ### Discord Embed Generator - URL: https://tolz.me/discord/embed-generator - Category: discord - Description: Build Discord embed JSON with title, description, color, fields, and footer. Fill in a title, description, color, author, footer, and up to 25 fields, then copy the JSON output. A live preview shows how the embed will look before you use it in a webhook or bot. #### FAQs **Q: What does the JSON output look like?** **A:** It is a JSON object with an embeds array containing one embed object. The embed has keys for title, description, color, author, footer, fields, and optionally timestamp. Paste the output under the embeds key in your webhook POST body. **Q: How do I add a Discord timestamp to an embed?** **A:** Enable Include current timestamp to add an ISO timestamp to the embed. Discord renders it as a relative countdown in the footer. For a fixed event time, generate a timestamp code separately and paste it into the description field manually. **Q: What is the maximum number of fields?** **A:** Discord supports up to 25 fields per embed. The form lets you add fields up to that limit. **Q: Can field values contain Discord markdown?** **A:** Yes. Field values support bold, italic, spoiler tags, inline code, and masked links. Headers with # do not render in field values. Code blocks work but the fences appear as text — inline backticks are more reliable for short snippets. **Q: Why does my webhook reject the JSON?** **A:** Check that the color is a number, not a quoted string. Make sure field names and values are strings, not empty. If the description exceeds 4096 characters, trim it. Discord also rejects embeds with both an empty title and empty description. **Q: Can I send an embed as a regular user?** **A:** No. Only bots and webhooks can send embeds. Regular Discord accounts can send plain text messages. If you need embed styling in a regular message, consider using ANSI color blocks instead. **Q: Is this tool free?** **A:** Yes. The form and JSON builder run entirely in your browser. ### Discord Webhook Sender & Message Builder - URL: https://tolz.me/discord/webhook-sender - Category: discord - Description: Build a Discord webhook message with embeds, files, and components v2, preview it, then post it straight from your browser. Write the message, fill in the embeds, attach files, and see exactly how Discord will draw it before you send. The webhook link stays in your browser and goes straight to Discord, so nothing about it is stored on Tolz.me. Send a new message, edit one you already posted, or load an old message back into the editor. #### FAQs **Q: Is this a webhook sender or only a message builder?** **A:** Both. The form builds the message and the preview shows the result, but the Send button posts the message to Discord. Edit updates a message you already sent, and Load message pulls one back into the editor. **Q: Can webhooks send interactive buttons?** **A:** A plain webhook can only send link buttons, which open a page when someone clicks them. Buttons that trigger a command need a webhook owned by an application with the right permissions. **Q: Can this tool schedule a message for later?** **A:** No. The sender posts the moment you press Send. For scheduled posts, run a script on a cron job or a task scheduler that calls the Discord webhook API. **Q: How do I delete a webhook?** **A:** Open Server Settings, choose Integrations, then Webhooks, and delete the webhook next to its name. The webhook info tool shows which channel a link posts into and how to rotate it if you are not sure which one to remove. **Q: What are the Discord webhook rate limits?** **A:** A webhook can make about 5 requests every 2 seconds. Go faster and Discord replies with 429 and a retry_after value. The sender shows that wait time so you know how long to pause. **Q: Why did Discord reject my message?** **A:** Most rejections come from a limit: content over 2000 characters, more than 10 embeds, a title over 256 characters, or a total embed character count above 6000. The counters in the editor turn red before you send, and the error panel lists the exact field Discord complained about. **Q: Can I edit a message I already sent?** **A:** Yes. Copy the message ID from the result panel, paste it into the message ID field, change the message, and press Edit. Discord replaces the original message with the new one. **Q: What is components v2?** **A:** It is the newer message format built from blocks: text, sections, containers, media galleries, separators, and rows of link buttons. A components v2 message cannot also carry plain text, embeds, or files. **Q: Is my webhook link safe to paste into this tool?** **A:** The link never reaches Tolz.me. It stays in the page while you work, and it is only written to sessionStorage if you tick remember in this tab. It is never placed in saved drafts, exports, or share links. Treat the link as a secret anyway, because anyone who has it can post to that channel. **Q: Is this tool free?** **A:** Yes. There is no account and no payment. Everything runs in your browser. ### Discord Webhook Info & Deletion Guide - URL: https://tolz.me/discord/webhook-info - Category: discord - Description: Read a Discord webhook link to see its ID, channel, and server, then find out how to delete or rotate it. Paste a webhook link and check what it points at: the webhook ID, the channel it posts into, the server, and whether it belongs to an application. The page also explains how to delete a webhook, how to rotate one without downtime, and what a leaked link can do. #### FAQs **Q: How do I delete a Discord webhook?** **A:** Open Server Settings, choose Integrations, then Webhooks, and press the bin icon next to the webhook. The webhook is removed at once and every link that pointed at it stops working. **Q: Can I move a webhook to another channel?** **A:** No. A webhook is tied to one channel. Create a new webhook in the channel you want, switch your script to the new link, then delete the old webhook. **Q: What does the webhook ID tell me?** **A:** It identifies the webhook. It is a snowflake, so the timestamp part of it shows roughly when the webhook was created. The channel ID and server ID are separate numbers that Discord returns when you look the webhook up. **Q: Does the lookup show the channel name?** **A:** No. Discord returns the channel ID for a webhook, not the channel name. You can read a snowflake ID with the Discord ID lookup tool to get the creation date, and match the channel ID against your own server. **Q: Is my webhook link sent to Tolz.me?** **A:** No. The link stays in your browser and the lookup request goes straight to Discord. If you tick remember in this tab, the link is kept in sessionStorage and is cleared when you close the tab. **Q: What happens if someone gets my webhook link?** **A:** They can post messages to that one channel, with any name and avatar they choose. They cannot read the channel history or act as your account. Delete the webhook to cut the link off for good. **Q: How do I rotate a webhook without breaking my scripts?** **A:** Create the new webhook first, update every script and job that posts through the old link, wait a moment, then delete the old webhook. Deleting first leaves the scripts failing until you update them. **Q: Does this tool cost anything?** **A:** No. There is no account and no payment. The page runs in your browser. ### Discord Bot Invite URL Generator - URL: https://tolz.me/discord/bot-invite-generator - Category: discord - Description: Generate Discord bot OAuth invite links with custom permissions. Enter your bot's Application ID, check the permissions your bot needs, and copy an OAuth invite URL ready to share with server admins. #### FAQs **Q: Where do I find my bot's Application ID?** **A:** Go to discord.com/developers, click your application, and look at the General Information page. The Application ID is the same as the client ID used in OAuth URLs. **Q: What permissions should I request?** **A:** Only check the permissions your bot actually uses. If it reads messages and posts replies, check View Channel and Send Messages. If it manages roles, also check Manage Roles. Leave everything else unchecked. **Q: Why should I avoid requesting Administrator?** **A:** Administrator grants every permission, which most server admins will decline. Only request it if your bot genuinely manages the entire server. Minimum-privilege invites get accepted more often. **Q: Does the invite URL expose my bot token?** **A:** No. The URL only contains the Application ID and a permissions integer. The token is a separate secret stored in your bot's code and must never be shared publicly. **Q: How do I update permissions after generating the URL?** **A:** Regenerate the URL with the new permissions selected and share the new link. The old URL will still work with the original permissions — revoke it from the Developer Portal if you need to retire it. **Q: Can anyone use the invite URL to add my bot?** **A:** Yes for public bots. The URL encodes the permissions, not a destination server. Anyone with the link can add your bot to any server where they have Manage Server permission. **Q: Is this tool free?** **A:** Yes. URL generation runs in your browser with no account required. ### Discord Reaction Role Template - URL: https://tolz.me/discord/reaction-role-template - Category: discord - Description: Copy reaction role setup messages for self-assign emoji roles. Pick from four reaction role panel templates, edit the emoji and role names, and post the message. Then configure your reaction role bot to watch it. #### FAQs **Q: Do reaction roles work without a bot?** **A:** No. Discord has no built-in reaction roles. You need a bot such as Carl-bot, Dyno, or Reaction Roles to link emoji reactions to role assignments. The template tool generates the message text only. **Q: How do I connect the message to my bot?** **A:** Enable Developer Mode in Discord settings, right-click the posted message, and copy the message ID. In your bot's dashboard or command interface, add a reaction role entry mapping each emoji to its target role using that message ID. **Q: Can I use custom server emoji?** **A:** Yes, if your server has the emoji. Type the custom emoji name directly into the template text, then add the same emoji as reactions on the message. Nitro members can use custom emoji from any server they share. **Q: How many emoji roles should one message have?** **A:** Five to eight per message is readable on mobile. If you need more options, split them across two messages in the same channel and pin both. **Q: What is the difference between ping roles and color roles?** **A:** Ping roles grant notification access — members who react get @mentioned when you message that role. Color roles are cosmetic — they change the color of the member's name in the sidebar and nothing else. **Q: Why did a member get the wrong role or no role?** **A:** Most likely the emoji in the bot config does not match the reaction on the message exactly. Check for stray spaces or typos. Also confirm the bot has permission to assign the target role. **Q: Is this tool free?** **A:** Yes. Templates copy instantly with no login required. ### Discord AutoMod Rules Template - URL: https://tolz.me/discord/automod-rules-template - Category: discord - Description: Copy AutoMod rule templates for keywords, spam, and mention limits. Browse AutoMod rule setup notes for keyword filters, spam protection, mention limits, and link restrictions. Recreate each rule in Server Settings, AutoMod. #### FAQs **Q: Does this tool enable AutoMod for my server?** **A:** No. This tool generates guidance text. You build the actual rules yourself in Server Settings, AutoMod. Discord hosts the filtering on its end, so no bot is required for basic protection. **Q: Is AutoMod free?** **A:** Yes for standard community servers. Discord includes AutoMod without extra cost. Some advanced features or usage tiers may require a paid plan depending on Discord's current offerings. **Q: Can I block Discord invite links with AutoMod?** **A:** Yes. The link filtering template covers invite link blocking. Discord AutoMod has a built-in option to block or flag server invites. You can also allowlist specific channels where link posting is permitted. **Q: What actions can AutoMod take when a rule triggers?** **A:** You can set AutoMod to block the message, timeout the sender for a set duration, or simply alert your staff channel. Block and timeout require server Boost level 1 or higher. Alert-only works on all servers. **Q: How do I avoid blocking legitimate messages?** **A:** Start rules in alert-only mode and review the mod log for a few days before switching to block. Short keywords like "free" catch too many false positives. Use specific phrases instead of single common words. **Q: Why did a legitimate message get blocked?** **A:** Your keyword list is probably too broad or too generic. Edit the rule, check the mod log to see what triggered it, and either add exceptions for specific channels or replace the keyword with a longer phrase. **Q: Is this tool free?** **A:** Yes. All template text is free to copy and adapt. ### Discord Announcement Generator - URL: https://tolz.me/discord/announcement-generator - Category: discord - Description: Generate server announcement messages with optional Discord timestamps. Pick an announcement type, set a date and time for the timestamp, edit the message, and copy the result. Works for patch notes, maintenance windows, giveaways, and new channel introductions. #### FAQs **Q: How does the timestamp feature work?** **A:** Pick a date and time in the datetime field. The tool converts it to a Discord timestamp using the F format, which displays as a full date and time in each viewer's local timezone. The {timestamp} placeholder in the template is replaced with that code. **Q: Can I edit the template text before copying?** **A:** Yes. The textarea is fully editable. If you accidentally delete the {timestamp} placeholder, add it back where you want the time code to appear. You can also add Discord markdown like **bold**, *italic*, or [masked links](https://example.com). **Q: What templates have a timestamp placeholder?** **A:** Maintenance and Giveaway templates include {timestamp}. Patch notes and New channel templates do not include a timestamp slot since they do not reference a specific event time. **Q: Can I use a different timestamp format?** **A:** Replace {timestamp} with a code from the timestamp generator tool. Use t for short time, T for long time, d for short date, D for long date, F for full date-time, or R for a relative countdown. **Q: Does the tool send the announcement automatically?** **A:** No. This generates message text only. Copy the output and paste it into your announcements channel manually, or pass it to your bot's scheduled message system. **Q: Can I add role mentions?** **A:** Add <@&roleId> manually after copying if you need to ping a role. Make sure the channel permissions allow role mentions. Reserve pings for announcements that require immediate attention from the whole server. **Q: Is this tool free?** **A:** Yes. All templates and generation run in your browser with no account needed. ### Discord Table Generator - URL: https://tolz.me/discord/table-generator - Category: discord - Description: Create a monospace Discord table inside a code block and copy it without broken columns. Discord markdown tables do not render in chat, so a monospace text table inside a code block is the practical fix. Build the grid by hand or paste CSV and TSV data, choose borders and alignment, then copy the finished message. The preview and character counter show what members will see before you post. #### FAQs **Q: Does Discord support markdown tables?** **A:** No. Discord does not render pipe tables. If you paste a markdown table into chat, members see the raw pipes and dashes. A text table inside a code block is the workaround because the monospace font keeps columns aligned. **Q: Why does my table look crooked in Discord?** **A:** The table is probably outside a code block, so Discord is using a proportional font. Copy the full message from this tool, including the triple backticks. Wide emoji and CJK characters can still shift a column because their display width differs from Latin letters. **Q: How do I paste a table from Excel or Google Sheets?** **A:** Copy the cells from the spreadsheet, paste them into the CSV or TSV box, and choose Fill grid from pasted data. Tab-separated data from a spreadsheet works best. CSV files with quoted commas are also supported. **Q: What is the character limit for a Discord table?** **A:** Discord's standard message limit is 2,000 characters at the time of writing, and Nitro raises it to 4,000. The counter includes the code fences, spaces, and border characters, so wide tables with long cells reach the limit sooner. **Q: Can I use emoji in a Discord table?** **A:** Yes, but emoji and CJK characters take more horizontal space than Latin letters. The tool measures display width to compensate, but fonts and clients vary. Check the preview on mobile if exact alignment matters. **Q: Why does the table wrap on mobile?** **A:** Narrow screens wrap long code block lines instead of scrolling sideways. The copied text is still correct, but the visual grid breaks. Use fewer columns, shorter headers, or split a wide table into two messages. **Q: Can I put a table in a Discord embed?** **A:** Code blocks usually work in embed descriptions and field values, but not in embed titles or footers. Test the embed in a private channel first because clients and bot libraries can handle markdown differently. ### Discord Invisible Name Generator - URL: https://tolz.me/discord/invisible-name - Category: discord - Description: Copy blank characters and build an invisible Discord display name or nickname. An invisible Discord name uses blank Unicode characters that draw nothing on screen. Copy one of four common characters or build a blank name of a set length, then test it in a display name or server nickname. Usernames cannot be invisible because Discord handles only allow a-z, 0-9, _ and . #### FAQs **Q: Can I make my Discord username invisible?** **A:** No. Discord usernames are the @handles used to find accounts, and they only allow lowercase letters, numbers, underscores, and periods. Blank Unicode characters are not accepted there. **Q: Will an invisible display name work?** **A:** Sometimes. Display names and server nicknames may accept blank characters such as Hangul Filler or Braille Pattern Blank. Discord can change its filters, and each server can set its own naming rules, so test before relying on it. **Q: Which blank character works best on Discord?** **A:** Hangul Filler (U+3164) is the most common choice and often survives name filters. Braille Pattern Blank (U+2800) is another option. Zero-width space and invisible separator are more likely to be stripped, so try them as fallbacks. **Q: Why did my invisible name turn back to normal?** **A:** Discord or the server likely removed the character when saving, or the name was reset by a filter or moderator. Try a different blank character, use a shorter run, or check the server rules before trying again. **Q: Can people still mention me if my name is blank?** **A:** Yes. Mentions use your account and user ID, not the visible display name. A blank name does not make you unmentionable or anonymous in moderation logs. **Q: Is an invisible name against Discord rules?** **A:** Discord does not publish a blanket ban on blank display names, but individual servers often do. Do not use a blank name to impersonate staff, evade moderation, or make reports harder to handle. **Q: How long can an invisible display name be?** **A:** Discord display names and server nicknames are commonly capped at 32 characters at the time of writing. The builder caps the output at 32 characters, but the limit can change. ### Discord Spoiler Text Generator - URL: https://tolz.me/discord/spoiler-text-generator - Category: discord - Description: Wrap Discord text in spoiler tags as one block, word by word, or letter by letter. Discord spoilers hide text behind a bar until someone clicks or taps it. Add the double-pipe syntax to a whole message, each word, or every letter, then copy the result. The character and pipe counters show how the extra tags affect Discord's message limit. #### FAQs **Q: What is the Discord spoiler syntax?** **A:** Wrap the hidden text in two pipe characters on each side, like ||this||. Discord shows a gray bar until the reader clicks or taps it. **Q: Why are my spoilers not working in Discord?** **A:** Check for a matching pair of double pipes, make sure the text is not inside a code block, and remove any pipe characters from the middle of the hidden text. Discord also does not handle nested spoilers reliably. **Q: Do spoiler pipes count toward the character limit?** **A:** Yes. Every pipe is a character in the message. Whole text mode adds four pipes, per-word mode adds four per word, and per-letter mode adds four per visible character. **Q: Can I spoil an image on Discord?** **A:** Yes. Attach the image and enable Mark as spoiler in the upload preview before sending. On mobile, select the image and use the spoiler option in the attachment preview. Bot and webhook uploads can prefix the filename with SPOILER_. **Q: How do I reveal a spoiler on mobile?** **A:** Tap the gray bar or the blurred image. If tapping does not work, try a long press or open the message in the desktop app. Gesture behavior can vary between app versions. **Q: Can I put a spoiler inside another spoiler?** **A:** Discord does not support nested spoilers in a reliable way. Keep each spoiler as a single layer and split the hidden text into separate sections if needed. **Q: Does per-letter spoiler mode always render correctly?** **A:** It produces adjacent spoiler tags, which most clients handle. Some updates can change how those tags parse. If the result looks wrong, per-word mode is more predictable and easier to read. ### Discord Small Text Generator - URL: https://tolz.me/discord/small-text-generator - Category: discord - Description: Create Discord subtext, Unicode small caps, superscript, and subscript from one input. Discord small text has two forms: subtext made with a special prefix, and Unicode characters that look like tiny letters. This tool generates both, plus superscript and subscript. Subtext works in messages, while the Unicode styles also work in display names, nicknames, and bios. #### FAQs **Q: How do I make small text in Discord?** **A:** Put -# at the start of a line, followed by a space, then your text. Discord renders that line as subtext. Every line needs its own prefix. **Q: Does subtext work in a Discord display name?** **A:** No. Subtext is a message format. In a display name or nickname field, the -# prefix appears as normal characters. Use Unicode small caps or superscript for profile text instead. **Q: Is this a real small font?** **A:** No. Small caps, superscript, and subscript use Unicode look-alike characters. Discord does not load custom fonts, so the result depends on the font installed on each device. **Q: Why do some letters stay normal in subscript?** **A:** Unicode does not include a subscript form for every letter. In this tool, b, c, d, f, g, q, w, y, and z stay full size. Superscript has better coverage but still has gaps. **Q: Can I use small caps in my Discord username?** **A:** No. Discord usernames only allow lowercase letters, numbers, underscores, and periods. Unicode small caps can go in your display name, server nickname, bio, and messages. **Q: Do screen readers read small caps normally?** **A:** Not always. Some screen readers read Unicode small caps letter by letter, skip them, or pronounce them as different characters. Use small text for secondary details, not rules or warnings. **Q: Why does -# show up as plain text?** **A:** The prefix must be at the very start of the line with a space after it, and it must not be inside a code block. Retype the hyphen if you copied it from a word processor that uses a different dash character. ### Roblox Display Name Checker & Generator - URL: https://tolz.me/gaming/roblox-display-name-checker - Category: gaming - Description: Check a Roblox display name against the published rules and generate aesthetic, cool, cute, or funny name ideas. Roblox display names follow a short set of length and character rules, and this checker shows which ones your name passes before you open Roblox. The generator builds themed ideas from word lists, and every idea already fits the local checks. Display names are not unique, and Roblox's text filter still has the final say. #### FAQs **Q: What are the Roblox display name rules?** **A:** At the time of writing, a display name must be 3 to 20 characters and can use letters, numbers, spaces, and underscores. It should not start or end with a space. Roblox also runs a text filter when you save the name. **Q: Are Roblox display names unique?** **A:** No. Several players can use the same display name. Usernames are the unique account handles, and they follow different rules, so there is no availability check for a display name. **Q: How often can I change my Roblox display name?** **A:** Roblox allows one display name change every seven days. This tool cannot see when you last changed yours, and it does not handle any fee that may apply. **Q: Why does Roblox reject a name that passes this checker?** **A:** The checker cannot run Roblox's text filter. It only tests length, characters, and spacing. A name can follow those rules and still break Roblox's community rules, so Roblox makes the final decision when you save it. **Q: Can I use symbols or aesthetic fonts in a Roblox display name?** **A:** Symbols, emoji, accented letters, and Unicode lookalike fonts are generally rejected. The checker flags them because they are outside the basic Latin letters, numbers, spaces, and underscores that display names normally allow. **Q: What is the difference between a Roblox display name and username?** **A:** A display name is the public label shown in games and chat, and it does not have to be unique. A username is the unique login handle. Usernames are 3 to 20 characters, use letters and numbers with one underscore, and the underscore cannot sit at the start or end. **Q: Does the generator check whether a name is available?** **A:** No. Display names are not unique, so there is nothing to check. The generator only makes sure its ideas fit the local length and character rules. Roblox can still filter an idea. ### Minecraft Color Code Generator - URL: https://tolz.me/gaming/minecraft-color-codes - Category: gaming - Description: Insert Minecraft color and formatting codes, preview them on a dark background, and copy Java, plugin, or MOTD output. Minecraft color codes are section-sign characters that change how text looks in the game. This generator inserts each code at your cursor, previews the result, and gives you the same string in section-sign form, ampersand form, or a server.properties MOTD line. A Bedrock switch adds Minecoin Gold and explains where the two editions differ. #### FAQs **Q: What does a Minecraft color code look like?** **A:** It is a section sign followed by one character, such as §6 for gold, §l for bold, or §r for reset. The game reads the code and draws the following text in that style. **Q: Do Minecraft color codes work in chat?** **A:** In vanilla Java chat, typed section signs usually appear as plain characters instead of colored text. The codes work in places such as the server list MOTD, written books, command output, and messages sent by plugins. Server software decides what is supported. **Q: What is the difference between Java and Bedrock color codes?** **A:** Both editions share §0 through §f and common formatting codes. Java uses §m for strikethrough and §n for underline. On Bedrock those letters are material colors, and Bedrock adds §g for Minecoin Gold. **Q: How do I use the MOTD output in server.properties?** **A:** Paste the MOTD output after motd= in server.properties, save the file, and restart or reload the server. The output writes the section sign as § because Java properties files expect that escape. **Q: Why does §g not work on Java?** **A:** Minecoin Gold is a Bedrock color. Java clients do not have that legacy code, so the text may lose its color or show the code literally depending on where it is used. **Q: Can I use hex colors in Minecraft?** **A:** Java Edition supports RGB colors in text components and in the expanded §x section-code form. Bedrock does not support arbitrary hex colors the same way. Plugin support varies, so check the documentation for the software you use. **Q: Why do §m and §n look different on Bedrock?** **A:** On Bedrock, §m and §n are material colors rather than strikethrough and underline. Switch the editor to Bedrock to see the material colors, and switch back to Java for the text decorations. ### Steam Artwork Showcase Cropper - URL: https://tolz.me/gaming/steam-artwork-cropper - Category: gaming - Description: Crop one still image into the wide main panel and narrow side panel used by a Steam artwork showcase, then download both PNGs. A Steam artwork showcase pairs a wide main image with a narrow side strip, and the gap between them is part of the layout. This cropper lets you frame one source image, set the gutter and height, and export two PNGs sized 506 and 100 pixels wide. Everything runs in your browser, and animated GIFs are not supported yet. #### FAQs **Q: What size is the Steam artwork showcase?** **A:** In the common layout, the main panel is 506 pixels wide and the side panel is 100 pixels wide. Both share a height, often around 506 pixels, with a gap of roughly 8 to 14 pixels. Steam layouts can change, so verify the result on your profile. **Q: Which image do I upload first?** **A:** Upload the wide main panel first and the narrow side panel second in the usual showcase editor. Steam can change the order controls, so check the saved profile and reorder the images if they appear swapped. **Q: Are my images uploaded to a server?** **A:** No. The image is loaded into a canvas in your browser, cropped locally, and downloaded as PNG files. No account connection or server upload is involved. **Q: Can I use a GIF?** **A:** Not yet. This tool exports static PNGs, so an animated GIF would lose its animation. Upload a PNG, JPG, or WebP still image instead. **Q: Why do the panels look misaligned on my profile?** **A:** The gutter or upload order probably does not match the layout. Try a gutter between 8 and 14 pixels, confirm that both files came from the same crop, and check that the main panel is in the wide slot. **Q: Does this work for a Workshop showcase?** **A:** This cropper targets the profile Artwork Showcase layout. A Workshop showcase or a custom profile theme may frame images differently, so compare the result on your own page before relying on it. ### Instagram Line Break Generator - URL: https://tolz.me/social/instagram-line-break-generator - Category: social - Description: Add blank lines to Instagram captions and bios with invisible spacer characters, then copy the spaced text in one click. Instagram can remove empty lines when text is pasted or saved, which is why a caption that looked spaced in your notes app can collapse into one block. This generator trims trailing spaces, places a Braille Pattern Blank or another spacer on each empty line, and counts the result against the 2,200 character caption limit and the 150 character bio limit. The text never leaves your browser, and you can copy the finished version without an account. #### FAQs **Q: Why do my Instagram line breaks disappear?** **A:** A blank line contains no visible characters, so some editors and apps trim it when text is saved or pasted. Behavior differs between captions and bios and can change between app versions. A spacer character gives the empty line something to hold so it is less likely to collapse. **Q: What is a Braille Pattern Blank?** **A:** It is the Unicode character U+2800. It looks empty on screen, but it is a real character, so Instagram and other apps treat the line as non-empty. It is the default blank line character in this tool. **Q: Does the spacer count toward the character limit?** **A:** Yes. Each spacer is one character, and line breaks may also count. The counter in the tool shows the total for the field you selected, so a text that fits before spacing can go over the limit after spacers are added. **Q: Will an invisible character work in an Instagram bio?** **A:** Often, but not always. The bio field is short and some app versions handle whitespace differently. Test the result in a draft before you publish, and try the visible dot if the invisible character is removed. **Q: Do screen readers read the blank character?** **A:** Some do. A screen reader may announce the Braille character, skip it, or pause. If the gaps carry meaning, add a visible separator or real punctuation so the text still makes sense without them. **Q: Can I use a visible dot instead?** **A:** Yes. Pick the visible dot option if you want the spacing to be obvious or if invisible characters keep disappearing. The dot counts as a character and will show in the caption. **Q: Is the Instagram caption limit 2,200 characters?** **A:** At the time of writing, captions allow 2,200 characters and bios allow 150. Instagram can change these limits, and the app's own counter has the final say. This tool uses those numbers as a close guide. ### YouTube Timestamp Link Generator - URL: https://tolz.me/social/youtube-timestamp-link-generator - Category: social - Description: Turn any YouTube link into a timestamped URL and check a chapter list against YouTube's description rules. A timestamp link starts a YouTube video at a chosen second, which is useful when you want to point someone to one answer, clip, or step instead of the whole video. Paste a watch, youtu.be, Shorts, live, or embed URL and this tool builds short, standard, and embed links. The chapter helper checks for a 0:00 start, at least three timestamps, ascending times, and chapters of at least 10 seconds. #### FAQs **Q: What is a YouTube timestamp link?** **A:** It is a normal YouTube URL with a time parameter. The player opens at that second instead of the start. The video itself is unchanged and still runs to the end. **Q: How do I make a YouTube link start at a specific time?** **A:** Add t=SECONDS to a youtu.be or watch link, or start=SECONDS to an embed URL. This tool converts a time such as 1:30 to 90 seconds and writes all three forms for you. **Q: Should I use t= or start=?** **A:** Use t= for links you paste into chat, social posts, or a website. Use start= for an embedded player. The number is the same in both cases. **Q: Do timestamp links work with YouTube Shorts?** **A:** The parser accepts Shorts URLs, but the Shorts player may ignore the start parameter on some devices. If the start time matters, share the standard watch link and test it on the target device. **Q: What are YouTube's chapter rules?** **A:** At the time of writing, the first timestamp must be 0:00, the description needs at least three timestamps, each chapter must last at least 10 seconds, and the timestamps must be in order. YouTube can change how it detects chapters. **Q: Why did YouTube not create chapters?** **A:** One of the rules is probably failing. Check the first timestamp, the total count, the spacing between timestamps, and whether each line follows the timestamp-and-title format. YouTube may also take time to process a new description. **Q: Does the tool store my video URLs?** **A:** No. Parsing and link building happen in your browser. Nothing is uploaded, and the tool does not connect to a YouTube account. ### TikTok Username Generator - URL: https://tolz.me/social/tiktok-username-generator - Category: social - Description: Generate TikTok username ideas in five styles and check a name against the basic format rules. TikTok usernames need to be short, easy to read, and within the platform's format rules. This generator builds aesthetic, funny, gaming, business, and personal name ideas from your keyword, then checks a username for allowed characters, length, and a trailing period. It cannot check availability, because TikTok does not provide a public lookup for outside tools. #### FAQs **Q: How do I use the generated TikTok usernames?** **A:** Tap an idea to copy it, then open TikTok and try it in your profile settings. The generator only creates text. It does not reserve the name or connect to your account. **Q: What characters can a TikTok username contain?** **A:** As commonly documented, usernames can use letters, numbers, underscores, and periods. Spaces, dashes, emoji, and most symbols are not accepted. The checker flags anything outside that set. **Q: How long can a TikTok username be?** **A:** The usual limit is 24 characters, and TikTok typically requires at least two. The tool uses 2 to 24 as its range, but the app's own error message is the final authority. **Q: Can this tool check whether a username is available?** **A:** No. TikTok does not offer a public username lookup, so no outside tool can confirm availability reliably. Try the name in the app or on the website to see whether it is free. **Q: What does the with your name style do?** **A:** It adds common personal prefixes and suffixes such as its, real, the, daily, and official to your keyword. It works best when the keyword is your name or the name you want on the account. **Q: Why does a name I like not appear in the list?** **A:** Every idea is filtered for length, allowed characters, and a clean ending. If a pattern would exceed 24 characters, it is removed. Shorten the keyword or try another style to get more options. **Q: Can I use a generated name on another platform?** **A:** You can, but check that platform's rules first. Instagram, Discord, and Roblox all have different character and length limits, so a name that fits TikTok may not fit somewhere else. ### Cron Expression Translator - URL: https://tolz.me/developer/cron-expression-translator - Category: developer - Description: A cron expression translator that reads a crontab line in plain English, breaks down every field, and previews the next ten run times. A cron expression translator turns a crontab line into a sentence you can check at a glance. Paste a five-field schedule and it explains what each field does, warns you about the cases where standard cron does something unexpected, and lists the next ten run times in your browser's timezone. Macros such as @daily and @reboot are handled too. #### FAQs **Q: How do I read a cron expression?** **A:** The five fields always run in the same order: minute, hour, day of month, month, day of week. The job runs when all five match the current time. Each field accepts an asterisk for every value, a number, a range like 1-5, a list like 0,30,45, or a step like */15. **Q: What does */5 mean in a cron expression?** **A:** It means every fifth value of the field it sits in. In the minute field, */5 runs on minutes 0, 5, 10, 15 and so on, which is the standard way to schedule a job every five minutes. **Q: Why does my cron job run on days I did not expect?** **A:** When both day-of-month and day-of-week are restricted, standard cron runs the job when either one matches instead of when both match. That Vixie cron rule surprises most people. The translator flags this case every time it applies. **Q: What is the difference between 0 and 7 in the day-of-week field?** **A:** Both mean Sunday. The field runs from 0 to 7 where 0 and 7 are Sunday, 1 is Monday, and 6 is Saturday. **Q: Does this tool handle Quartz or AWS EventBridge schedules?** **A:** No. It reads the standard five-field format. Quartz and EventBridge use six fields with seconds first and a question mark for an empty day field, so those expressions are rejected with a field count message. **Q: Which timezone do the next run times use?** **A:** The timezone of your browser. Cron on a server uses that server's system timezone, which is often UTC, so compare the two before you rely on the list. **Q: Is this cron translator free?** **A:** Yes. It runs in your browser with no account, upload, or payment required. ### Invisible Character Remover - URL: https://tolz.me/text/invisible-character-remover - Category: text - Description: Find hidden characters such as zero width spaces, soft hyphens and directional marks, and remove them with one click. An invisible character remover finds the Unicode characters that draw nothing inside your text, names each one, and takes it out. Paste in a sentence and you get a list with the name, the code point and the count for every hidden character found, plus a view that shows exactly where each one sits. It also converts unusual spaces such as the no-break space into normal ones. If you want the opposite trick, the invisible name tool does the same job in reverse. #### FAQs **Q: What does an invisible character remover actually do?** **A:** It scans every character in your text against a list of Unicode code points that render as nothing, then reports how many of each one it found. You choose which characters to remove, and the tool gives you the cleaned text to copy. **Q: Where do hidden characters come from?** **A:** Most come from a copy and paste. Search engines put zero width spaces inside phrases, word processors mark optional hyphens with a soft hyphen, translation tools leave directional marks behind, and older systems add a byte order mark to exported text. **Q: Will removing hidden characters make my writing pass an AI detector?** **A:** No. Detectors look at phrasing, structure and word choice. A character that draws nothing plays no part in that judgement, so cleaning the text will not change the score you get. **Q: What is the narrow no-break space that shows up in some AI output?** **A:** It is U+202F, a space that stops a line from breaking. Some people asked OpenAI whether it was a hidden marker. At the time of writing, OpenAI said in its community forum that the narrow no-break spaces seen in some o3 and o4-mini output were not a watermark. **Q: Which hidden characters should I keep?** **A:** Keep the zero width joiner when it holds an emoji sequence together, or a family emoji splits into separate pieces. A soft hyphen can also be worth keeping when the word is meant to break across a line. Every character has its own checkbox, so the choice is yours. **Q: Does the tool upload my text?** **A:** No. All the checking and cleaning happens in your browser, so the text never leaves your device and nothing is stored. **Q: Will removing a zero width space change what my text means?** **A:** Usually not. In most cases it removes an invisible gap that was never meant to be read. The exception is text where the character is deliberate, such as a search engine keyword variation, so read the highlighted view before you copy. ### Chmod Calculator - URL: https://tolz.me/developer/chmod-calculator - Category: developer - Description: Convert file permissions between octal, symbolic, and ls -l forms, including setuid, setgid, and the sticky bit. A chmod calculator converts file permissions between the octal number, the symbolic string, and the letters you see in an ls -l listing. Tick the owner, group, and others boxes for read, write, and execute, or paste a mode you already have. Setuid, setgid, and the sticky bit are included, and the tool outputs a chmod command you can paste into a terminal. #### FAQs **Q: What does chmod 755 mean?** **A:** In octal, 7 is read plus write plus execute for the owner, 5 is read plus execute for the group, and the last 5 is read plus execute for everyone else. The symbolic form is rwxr-xr-x. **Q: What is the difference between 644 and 755?** **A:** 644 is rw-r--r--, so everyone can read the file but only the owner can change it. 755 is rwxr-xr-x, so everyone can also run it. Directories need the execute bit for people to enter them, which is why 755 is the usual directory mode. **Q: What does the sticky bit do?** **A:** On a directory, the sticky bit lets everyone create files but stops one user from deleting another user's files. That is why /tmp is 1777. It appears as a t in the symbolic form, or as the leading 1 in the octal form. **Q: What are setuid and setgid?** **A:** Setuid on an executable makes it run with the privileges of the file owner, which is how passwd works. Setgid on a directory makes new files inherit the directory's group. In symbolic form they show up as s in the execute position, and in octal as a leading 4 and 2. **Q: How do I convert an ls -l line such as drwxr-xr-x into a number?** **A:** Paste it into the ls -l field. The leading d marks a directory and is ignored. The tool converts the remaining nine characters into the octal form, which for drwxr-xr-x is 755. **Q: Is chmod 777 safe to use?** **A:** Not on a server. At 777 every account on the machine can change the file, and at 666 every account can read and write it. Use 755 for public files and directories, and 600 for anything secret. **Q: Is this chmod calculator free?** **A:** Yes. It runs in your browser and never touches your files or runs a command. ### Em Dash Remover - URL: https://tolz.me/text/em-dash-remover - Category: text - Description: Replace every em dash in your text with a comma, semicolon, colon, hyphen or your own text, with a side by side preview. Remove em dashes from text in one pass and choose what takes their place. The tool swaps the em dash for a comma, a semicolon, a colon, a hyphen, a spaced hyphen, nothing at all, or any text you type, then tidies the spacing around it. A side by side panel shows the original and the rewritten version so you can check every change before you copy it. #### FAQs **Q: What is the difference between an em dash and an en dash?** **A:** The em dash (U+2014) is the long one and marks a break inside a sentence. The en dash (U+2013) is shorter and marks a range or a span, such as 2013–2015. The hyphen (U+002D) is the short one and joins words such as well-known. **Q: What should I replace an em dash with?** **A:** It depends on the sentence. A comma suits a dash that continues the same thought, a semicolon suits two complete statements, and a colon suits a dash that introduces an explanation or a list. The tool shows the result for every dash so you can judge each one. **Q: Will removing em dashes make my writing look more human?** **A:** No. Punctuation is not what a detector looks at, and swapping a dash for a comma does not change the rhythm or the phrasing of a sentence. Use the tool to clean up a house style, not to fool a detector. **Q: How do I remove em dashes from a long text?** **A:** Paste the whole text into the box, pick the replacement you want, and copy the result. The counters show how many em dashes, en dashes and double hyphens were found and replaced. **Q: Why is there a double hyphen option?** **A:** An em dash typed as plain text often turns into two hyphens. The option treats any run of two or more hyphens as a dash, which is useful for old documents. Leave it off for text where a double hyphen is a real compound, such as well--known. **Q: Is my text uploaded to a server?** **A:** No. The replacement runs in your browser, so the text stays on your device and nothing is stored. **Q: Why does one sentence read badly after the change?** **A:** That dash was doing more than joining two clauses. Keep the dash in that sentence, or use the custom option to give it a different mark. A mechanical swap cannot know the difference between a dash that carries meaning and one that does not. ### Unix Timestamp Converter - URL: https://tolz.me/developer/unix-timestamp-converter - Category: developer - Description: Convert Unix timestamps to dates and back, with automatic unit detection for seconds, milliseconds, microseconds, and nanoseconds. A Unix timestamp converter turns a number of seconds since 1970 into a date, and a date back into a number. The tool detects whether the value is in seconds, milliseconds, microseconds, or nanoseconds, then shows UTC, your local timezone, ISO 8601, and a relative phrase such as in 3 hours. A live counter gives you the current timestamp, and there are snippets for JavaScript, Python, bash, and SQL. #### FAQs **Q: What is a Unix timestamp?** **A:** It is the number of seconds since 1 January 1970 at midnight UTC. A timestamp identifies one exact instant and carries no timezone of its own, which is why it stores and compares reliably. **Q: How do I convert a Unix timestamp to a date?** **A:** Paste the number into the timestamp field. The converter reads it as seconds, milliseconds, microseconds, or nanoseconds based on the number of digits, and shows the UTC time, your local time, the ISO 8601 string, and a relative phrase. **Q: How can I tell seconds from milliseconds?** **A:** By the number of digits. Ten digits is seconds, thirteen is milliseconds, sixteen is microseconds, and nineteen is nanoseconds. You can also pick the unit yourself from the dropdown if the automatic guess is wrong. **Q: What is the year 2038 problem?** **A:** A timestamp stored in a signed 32-bit integer runs out of room on 19 January 2038 at 03:14:07 UTC. The value wraps around and the system reports a date in 1901, or the program refuses to handle it. Use a 64-bit column in anything you build now. **Q: How do I get the current Unix timestamp in code?** **A:** In JavaScript use Math.floor(Date.now() / 1000), in Python int(time.time()), in bash date +%s, and in SQL SELECT EXTRACT(EPOCH FROM NOW()). The snippets panel in the tool copies each one. **Q: Are Unix timestamps affected by my timezone?** **A:** No. The number itself is the same everywhere. Timezones only change the local text shown next to it, which is why a timestamp stored in UTC reads the same for every user. **Q: Is this unix timestamp converter free?** **A:** Yes. It runs in your browser with no account needed, and nothing you paste is sent anywhere. ### Sort Lines Online - URL: https://tolz.me/text/sort-lines - Category: text - Description: Sort lines alphabetically, in natural number order or by length, and remove duplicate or blank lines. Sort lines online without opening a spreadsheet. Paste a list, choose A to Z, Z to A, natural order, or sort by line length, then remove duplicates, trim the blanks and drop the empty lines. The counters show how many lines went in and how many came out, and the result is plain text that pastes straight back where it came from. #### FAQs **Q: How do I sort lines alphabetically?** **A:** Paste one item per line into the box, leave the sort order on A to Z, and copy the result. Sorting uses the collation rules of your browser, so accented letters sit next to the letters they belong with. **Q: What is the difference between A to Z and natural order?** **A:** A to Z compares numbers as text, so item 10 comes before item 2. Natural order reads them as numbers, so item 2 comes first. Use natural order for numbered lists, versions and file names. **Q: How do I sort lines by length?** **A:** Pick Shortest first or Longest first in the sort order. Lines of the same length fall back to alphabetical order, which is handy for spotting the one line that does not belong. **Q: Can I remove duplicate lines?** **A:** Yes. Turn on Remove duplicate lines and the first copy of each line is kept while the rest are dropped. By default two lines count as duplicates only when they match exactly, and you can switch that off to treat Apple and apple as the same line. **Q: Does the sort keep capital letters in a separate place?** **A:** It depends on the case setting. With case ignored, Banana and banana land next to each other, which is what most people want for a list of names. With case respected, the capital letter sorts first, the way the Unicode collation rules treat it. **Q: Is the text uploaded to a server?** **A:** No. Sorting happens in your browser, so the list stays on your device and nothing is stored. **Q: Why is there one more line in the output than I expected?** **A:** A file that ends with a line break has an empty line after the last item. Turn on Remove empty lines and the count matches what you see. ### Hours Between Two Times Calculator - URL: https://tolz.me/calculator/hours-between-times - Category: calculator - Description: Work out the hours between two clock times, subtract unpaid breaks, and total a whole week in hours, minutes, and decimal hours. This hours between two times calculator gives you the exact gap between a start and an end time, including shifts that run past midnight, unpaid breaks, and a full timesheet total. Results come in hours and minutes, in decimal hours such as 7.75, and as an optional pay estimate at your own hourly rate. Everything is worked out in your browser, so none of your shift times are uploaded. #### FAQs **Q: How do I calculate the hours between two times?** **A:** Type your start and end time in the first row. The tool subtracts the end from the start, adds 24 hours when the end time is earlier so an overnight shift still counts, and then removes any unpaid break you enter. The answer shows next to the row as hours and minutes, and in the total cards below. **Q: How do I convert 7 hours 45 minutes to decimal hours?** **A:** Divide the minutes by 60 and add them to the hours. 45 divided by 60 is 0.75, so 7 hours 45 minutes is 7.75 decimal hours. The tool shows 7.75 for that shift, and 7.50 for a seven and a half hour day. **Q: How do I work out an overnight shift?** **A:** Put the evening start time and the morning end time in the same row. When the end time is earlier than the start time the tool treats it as the next day, so 22:00 to 06:00 comes to 8:00. A badge on the row shows that the overnight rule was applied. **Q: Does the calculator subtract an unpaid lunch break?** **A:** Yes. Type the length of the break in minutes in the break box on that row and it comes off the shift length. A 30-minute break on a 09:00 to 17:00 shift turns 8:00 into 7:30. **Q: How do I total a whole week of hours?** **A:** Use the Add day button to create a row for each day, then fill in the times. The total card adds every row together, and the Copy summary button copies the day by day breakdown with the total and the pay estimate. **Q: How is the decimal number rounded?** **A:** Decimal hours are rounded to two decimal places, so 8 hours 20 minutes shows as 8.33. The hours and minutes value stays exact to the minute, and the pay estimate is calculated from the exact time and rounded once at the end. **Q: Is the pay figure the same as what my employer pays?** **A:** Not necessarily. The estimate multiplies your worked hours by the rate you type. It does not apply tax, overtime rules, or the way your contract treats breaks, and those rules differ by employer and by country. Check the real figure with your payroll system. ### Aspect Ratio Calculator - URL: https://tolz.me/calculator/aspect-ratio-calculator - Category: calculator - Description: Calculate the missing side of an aspect ratio, reduce any pixel size to its simplest ratio, and read the dimensions of an image. This aspect ratio calculator works with pixels in both directions. Give it a ratio such as 16:9 and one side to get the other, or enter a pixel size to see its simplest ratio and decimal value. Drop an image onto the page to read its dimensions without uploading it, and pick from common sizes such as 1920 by 1080 or A4 at 300 DPI. Everything runs in your browser. #### FAQs **Q: How do I calculate an aspect ratio from pixels?** **A:** Reduce both numbers by their largest common divisor. For 1920 by 1080 the largest shared divisor is 120, so the simple ratio is 16:9. The tool does that reduction for you and shows the decimal form as well, with 1920 divided by 1080 rounded to 1.7778. **Q: How do I find the missing width for a 16:9 video at 1080 pixels high?** **A:** Multiply the height by the ratio: 1080 times 16 divided by 9 gives 1920. Type 16:9 in the ratio box, set the height to 1080, and the width fills in as 1920 pixels. **Q: What is the decimal aspect ratio of 16:9?** **A:** 16 divided by 9 is 1.7777..., which the tool shows as 1.7778 after rounding to four decimal places. The decimal form is useful for software that expects a single number rather than a ratio. **Q: Can an aspect ratio calculator read the size of my own image?** **A:** Yes. Drop a file onto the upload area and the tool reads its pixel dimensions in your browser, then fills in the width, the height, and the simplified ratio. The file is not uploaded, so a private photo stays on your device. **Q: Does resizing to the same aspect ratio stop a picture from stretching?** **A:** Yes. Scaling both sides by the same factor keeps the proportions. If you change the ratio, the program has to crop, stretch, or add bars, and each of those changes what you see in the frame. **Q: Which aspect ratio should I use for a social post?** **A:** It depends on the platform and where the post appears. Square posts use 1:1, portrait feeds commonly use 4:5, and wide banners use 16:9 or wider. Platforms adjust their recommendations regularly, so check the current spec before you build a whole set of assets. **Q: Why does 1024 by 768 show up as 4:3?** **A:** Because the tool reduces a size to its simplest whole number pair. 1024 and 768 share a divisor of 256, and 1024 divided by 256 is 4 while 768 divided by 256 is 3, so the shape is 4:3. ### PNG to ICO Converter (Favicon Maker) - URL: https://tolz.me/image/png-to-ico - Category: image - Description: Turn one PNG, JPG, WebP, or SVG into a multi-size favicon.ico file, plus the PNG sizes a website needs. A favicon.ico file holds several image sizes in one container, and browsers pick the one that fits the tab, the bookmark bar, or the shortcut. This PNG to ICO converter builds that file from any image you upload, with a crop or pad option for the square and your choice of sizes from 16 to 256 pixels. You also get the individual PNGs, including the 180 pixel apple-touch-icon and the 192 and 512 pixel icons a web app manifest uses, plus the link tags to paste into your HTML. Everything runs on your device, so the image is never uploaded. #### FAQs **Q: What image formats can I turn into a favicon.ico?** **A:** PNG, JPG, WebP, and SVG all work. The image is decoded by your browser and drawn onto a canvas, so any format your browser can display can be converted. **Q: Which favicon sizes should I include in the .ico file?** **A:** 16, 32, and 48 pixels cover the browser tab, the favorites list, and the Windows taskbar, and that is the set most sites use. Add 64, 128, and 256 when the same file also has to work as a Windows shortcut or an installer icon. **Q: Why do I need apple-touch-icon.png and manifest icons as well?** **A:** They serve different places. iPhone ignores the favicon and uses the 180 by 180 apple-touch-icon for the home screen shortcut. Chrome and Android read the 192 and 512 pixel PNGs from your web app manifest when someone installs your site. The .ico file still matters for Windows. **Q: Does the .ico file really contain an ICO, or is it a renamed PNG?** **A:** It is a real ICO container. The file has the six byte ICO header, one sixteen byte directory entry per size, and the image data after that. The entries themselves are PNG images, which Windows Vista and later and every current browser read. Very old software may need an uncompressed ICO instead. **Q: My new favicon does not show up. What should I check?** **A:** Start with caching. Browsers keep favicons for a long time and some hold on to the old file until you clear site data, so test in a private window first. Then check that the file is named favicon.ico and sits in the root folder of your site, and that the link tags are inside the head element. **Q: Is my image uploaded to a server?** **A:** No. The file is read with the browser's own file API, drawn on a canvas, and encoded into the .ico inside this tab. Nothing is sent anywhere and nothing is stored after you close the page. **Q: Why does my 16 by 16 icon look blurry while the 256 one looks fine?** **A:** A 16 pixel icon has 256 pixels in total, so fine detail cannot survive the trip down. Simplify the mark, drop thin strokes, and make sure it has a solid silhouette. If the source is a wide wordmark, crop to fill so a recognizable piece of it fills the small square. ### Color Palette from Image - URL: https://tolz.me/image/color-palette-from-image - Category: image - Description: Pull the dominant colors out of any image and copy them as HEX, RGB, or HSL, with the contrast of each one against white and black. A color palette from image is more useful than a list of random swatches, because the colors come from the picture you actually want to match. This tool extracts five to ten dominant colors, shows the hex, RGB, and HSL value of each one, and prints the WCAG contrast against white and black so you can pick readable text colors. An eyedropper reads a single exact pixel, and the palette exports as CSS variables, a Tailwind config snippet, or JSON. The image is processed in your browser and never uploaded. #### FAQs **Q: How does the tool decide which colors are dominant?** **A:** It shrinks the image to a small canvas, then groups similar pixels together with a k-means pass. Each cluster is averaged into one color, and the colors are sorted by how much of the picture they cover. The starting points are fixed, so the same image always gives the same palette. **Q: How many colors can I extract at once?** **A:** Between five and ten. Five is usually enough to describe a photo, and past ten the clusters start to describe individual objects rather than the overall look of the image. **Q: What do the contrast numbers mean?** **A:** They are the WCAG 2.1 contrast ratio of each color against white and against black. 4.5 to 1 is the minimum for normal body text, 3 to 1 covers large text and interface borders, and 7 to 1 is the stricter AAA level. Each swatch also tells you whether black or white text reads better on it. **Q: Can I export the palette for my project?** **A:** Yes. The export panel writes CSS custom properties, a Tailwind config snippet for Tailwind 3 with the matching @theme block for Tailwind 4, or a JSON list with the hex, RGB, HSL, contrast ratios, and the share of the image each color covers. **Q: Why is there a white swatch in a photo with no white in it?** **A:** Anti-aliased edges and soft shadows create pixels that are almost white or almost black, and those pixels are part of the image, so the palette picks them up. Use the eyedropper to find the color you actually wanted. **Q: Does the eyedropper read the exact pixel I click?** **A:** Yes, it reads the single pixel under the pointer and reports its hex, RGB, and HSL value with the same contrast numbers. With the preview focused, the arrow keys move the sample point and Enter samples it, so you do not need a mouse. **Q: Which file types work, and is my image uploaded anywhere?** **A:** PNG, JPG, WebP, GIF, and SVG are all supported, and the first frame is used for an animated GIF. The file is decoded and drawn on a canvas inside this tab, so nothing is sent to a server and nothing is kept once you close the page. ## Guides ### 50,000 Fake Statistics and ChatGPT Kept Citing Them: How to Check a Number Before You Use It - URL: https://tolz.me/blog/fake-statistics-ai-citations-how-to-check - Published: 2026-10-21 - Updated: 2026-10-21 - Tags: AI, Research, Fact Checking A site published 50,000 invented statistics in a day, and ChatGPT kept citing them. Here is how to check any number before you use it. In March 2026, Metehan Yesilyurt started building a site called stateglobe.com to test what happens when AI tools go looking for statistics. Every number on it was invented. Yesilyurt wrote it up on October 7, 2026, on the Peec AI blog, under the title Publish stats, get cited. The post is worth reading for the numbers in it, and worth reading even if you never plan anything like this, because it shows exactly how a false claim travels now. ## How it was built - 50,000 statistics pages generated in one day with GPT-4.1. - Every number on those pages invented and unchecked. - A cost of about $30 for the pages, plus roughly $20 more for 15,000 listicles. The idea behind it: as more articles get written with LLMs, those models go looking for numbers to fill gaps in what they know, and statistics pages are exactly what they search for. ## What happened next On the search side, the plan failed. Google's visibility collapsed after two updates, and the pages were deindexed. On the assistant side, it kept working. In the 30 days before the post, stateglobe.com received more than 600,000 bot requests. About 72,000 of them came from ChatGPT-User, which is the user agent ChatGPT uses when it fetches pages in real time to answer a question. Analytics on the site showed more than 10,000 sessions from ChatGPT since March. Ahrefs and Semrush showed more than 330 referring domains, including Shopify, DHL, and Bluehost. ## The invented numbers reached real publications Yesilyurt lists where the made-up numbers turned up: an academic paper, investment research from Qatar's largest financial group, an Indian economics publication, and a Canadian influencer community report. He notes these are only the citations that linked back to the site. The site now carries a disclaimer. ## Why one invented number travels so far A statistic page is the easiest kind of source for an assistant to use. It offers one number, one topic, and no argument to weigh. When a model needs a figure for a question, a page that contains exactly one figure fits perfectly, and nothing on the page contradicts it. Then the number leaves the page. It lands in an answer, the answer gets pasted into a document, and the document cites the page. At that point the claim has a chain of custody that looks ordinary, and the original page is still sitting there for anyone who follows the link. ## Retrieval is not ranking The part worth keeping from the write-up is his split between retrieval and ranking. The site lost its search visibility completely, and that barely dented the citation flow. Assistants pulled the pages because the pages answered the question, not because those pages ranked for it. Google rankings and being retrievable by an LLM, he argues, are two different problems. He also notes that LLM pipelines do not fact-check numbers. A number gets retrieved, quoted, and cited, and the check that would catch a false one happens somewhere else or not at all. He calls the result citation laundering. And he is plain about the lesson for himself: you should never publish fake content. ## How to check a statistic before you use it 1. Find the primary source. Search the exact number in quotes. If it only shows up on aggregator sites and news summaries, keep digging for the study, dataset, or filing underneath. 2. Read who published it. A number from a press release, a vendor survey, or a marketing page needs different handling than one from a peer-reviewed paper or a government dataset. 3. Check the date. Statistic pages get republished for years. A figure from 2019 sitting in 2026 prose may describe a market that no longer exists. 4. Ask what was measured and who was asked. Sample size, population, and collection method decide whether the number means what your sentence says it means. 5. Check that the source says what you are claiming. Percentage of what, out of how many, compared with which baseline. 6. Trace one step further. If you are citing a summary, open the summary's source. If that source points back to an aggregator, the chain ends there. One practical note on step one. Text copied from a web page or a chat window can carry invisible characters, which can make a quoted search miss. If a quoted search comes back empty, run the text through the [invisible character remover](https://tolz.me/text/invisible-character-remover) to see what the string really contains. ## What to check on the page itself - No named author and no publication date. - No methodology section, no sample description, no definition of what was measured. - One number per page, with no other figure to compare it against. - A figure with no unit, no population, and no time period attached. - Writing that reads like a summary of a summary and cites no source of its own. ## The same checks, applied to an AI answer 1. Open the citation and look for the number on the page. If it is not there, stop and treat the figure as unverified. 2. Look for the primary source behind the citation, not only the citation itself. 3. Check that the number in the answer matches the number in the source, including units and time period. 4. Find a second independent source before you publish, ideally one with a different method. ChatGPT's own behaviour is part of the story here. When it fetches a page in real time to answer a question, the answer can rest on a single page it read seconds earlier, with no second source to check it against. ## Warning signs in an AI answer - A precise number with no named source. - A link that turns out not to mention the number when you open it. - A figure that is unusually round, or unusually precise for the topic. - The same number appearing across several sources that all trace back to one origin. ## Keeping a source log Most false statistics reach a finished piece because nobody owned the verification step. A cheap fix is a short record kept next to the draft: the number as written, the source, the date you checked it, and the sentence in the source that supports it. That last field is the one people skip, and it is the one that saves you. If you cannot point at the line that supports a number in your own draft, you do not have a source yet. > Tip: Treat a number in an AI answer as a lead rather than a source. The citation tells you where to look. Reading the page is still your job. ## A rule for the draft One rule covers most cases: do not paste a number into a draft until you have opened the source. Keep candidate figures in a separate list with the link attached, and promote a number into the text only after you have read the part of the source that supports it. It slows the first minute and saves the last hour of edits. The same rule applies when a colleague sends you a number in a message. Ask for the source in that same thread, not three days later when the figure has already reached three decks. ## What to do when the number turns out to be fake Say so plainly and correct it. If you published it, fix the page and note the change. If somebody else published it, tell the author. A correction costs a paragraph. A statistic that spreads because you stayed quiet costs more. If the number came from an assistant, the correction also belongs in the thread where you first pasted it. That is where the copies started. ## How Tolz.me handles numbers We do not publish invented statistics, and we do not republish a number whose source we have not opened. When a guide here needs a figure, it comes from a source we link, with the publisher named. The [blog](https://tolz.me/blog) collects the rest of our guides. ## Sources - [Peec AI: Publish stats, get cited](https://peec.ai/blog/publish-stats-get-cited-the-tactic-that-put-my-data-all-over-the-web), by Metehan Yesilyurt, October 7, 2026 - [Hacker News discussion of the post](https://news.ycombinator.com/item?id=50165801) ### Agent Skills Are Trending on GitHub: What a SKILL.md File Is and How to Use One - URL: https://tolz.me/blog/agent-skills-skill-md-explained - Published: 2026-10-21 - Updated: 2026-10-21 - Tags: Agent Skills, SKILL.md, OpenCode, Developer Tools Skill and plugin repos filled GitHub's weekly trending list. Here is what a SKILL.md file contains, how agents load it, and how to install one. GitHub's trending page usually turns up frameworks, small libraries, and the occasional database engine. The week of October 21, 2026 was different. Several of the entries were skill and plugin repositories, including Anthropic's official Claude Code plugin directory, the Cursor plugin specification, Android's skills repo, and a science-focused library of agent skills. A SKILL.md file is a much smaller thing than the word plugin suggests, so it is worth ten minutes to understand what is actually trending and why. ## What showed up on the trending page The repos that landed on GitHub's weekly trending list during the week of October 21, 2026 each describe themselves differently: - anthropics/claude-plugins-official calls itself an "Official, Anthropic-managed directory of high quality Claude Code Plugins". - cursor/plugins holds the Cursor plugin specification and official plugins. - android/skills is Android's own skills repository. - K-Dense-AI/scientific-agent-skills describes itself as an Agent Skills library for science, compatible with Cursor, Claude Code, Codex, and other tools that follow the open Agent Skills standard. Read those descriptions together and a pattern shows up. Several vendors are publishing the same kind of unit: a folder holding instructions, designed to be picked up by a coding agent. The naming differs from tool to tool. The shape is the same, and the shape is small enough to learn in one sitting. ## What a skill actually is A skill is a folder with a SKILL.md file at its root. The file starts with YAML frontmatter, then the body of the skill. The frontmatter tells the agent what the skill is called and when it might be useful. The body tells the agent what to do. Here is a complete example: ``` --- name: release-notes description: Turn a list of merged pull requests into customer-facing release notes. Use when asked to write changelog entries or summarize what shipped this week. license: MIT metadata: version: "1.2.0" --- # Release notes ## Inputs - The list of merged pull requests for the target week - The previous release notes, for tone and format ## Steps 1. Group the pull requests by user-visible change, not by internal module. 2. Write one sentence per change. Lead with the effect on the user, not the implementation. 3. Drop anything a user would not notice, including refactors, test changes, and dependency bumps. 4. Keep the order stable: features first, fixes second. ## Output format - A short intro paragraph, no more than two sentences. - Bullet list, one bullet per change, past tense. - A "Known issues" section only when a listed item is still open. ``` Two fields are required: name and description. Everything else is optional. The name has rules that tools validate strictly, so it is worth getting right the first time: - 1 to 64 characters. - Lowercase letters and numbers separated by single hyphens. The pattern is `^[a-z0-9]+(-[a-z0-9]+)*$`, so `code-review` passes and `code_review` does not. - It has to match the folder name. A folder called `code-review` holding a skill named `code-review` is the arrangement that survives validation. The description can run from 1 to 1024 characters, and it is the part the agent reads before deciding whether to open the skill. Write it as a trigger rather than a summary. Say when to use the skill, in the words a user would actually type, because that is what the agent matches against. ### The optional fields - license: the license that covers the skill. - compatibility: a note about which tools or versions the skill targets. - metadata: a map of string keys to string values, such as a version. ## How an agent finds and loads a skill OpenCode's documentation describes the loading model plainly. The agent is given the list of skill names and descriptions. It does not pull the body of every installed skill into context. When a task matches one, the agent calls a skill tool and loads the full content on demand. That design is what makes a large library cheap. Twenty installed skills cost twenty lines of names and descriptions until one of them is actually needed. OpenCode searches six locations: - `.opencode/skills` inside the project. - `~/.config/opencode/skills` for the current user. - `.claude/skills` inside the project. - `~/.claude/skills` for the current user. - `.agents/skills` inside the project. - `~/.agents/skills` for the current user. Four of those paths are not OpenCode specific. `.claude/skills` and `.agents/skills`, in the project and in your home folder, are the Claude-compatible and agent-compatible locations, so a skill kept there is not locked to OpenCode. The same open Agent Skills format is what the science library above says it follows. Permissions live in `opencode.json`, and they can allow, deny, or ask per skill name pattern. A denied skill is hidden from the agent, and an ask rule makes the agent get your approval before it loads that skill. ## Installing skills from GitHub Writing a skill by hand takes a minute. Installing one takes two commands: ``` npx skills find release notes npx skills add owner/repo ``` The `skills find` command searches the catalog hosted at skills.sh, and `skills add` pulls the skill from a GitHub repository into the place your agent looks for skills. Because the repository is where the skill lives, it is also where you read it before installing. ## Writing a skill that holds up 1. Give it one job. A skill that covers three tasks becomes a document the agent half-follows. 2. Write the description as a trigger, and include the words a user would actually type. 3. List the inputs first, then the steps in order, so the agent does not have to guess what to collect. 4. Show the output format. A worked example removes more guesswork than another paragraph of prose. 5. Keep the body short enough to read in one pass. If it runs long, split the reference material into a separate file the skill points to. The example above follows that shape: inputs, numbered steps, output format. Nothing else. A skill that a person can follow on a first read is a skill an agent can follow too. ## A skill runs with the agent's permissions This is the part worth slowing down on. A skill is instructions the agent will follow, and it runs with whatever the agent can already do. If the agent can read your files, run shell commands, and call your API, then a skill can tell it to do those things. The skill does not get its own sandbox. Before you install one: - Read the SKILL.md first, and every file in the folder including any script it references. - Look for anything that reaches for credentials. A skill that asks for an API key, a token, or a password is a reason to walk away. - Treat the shell commands and network calls inside a skill the way you would treat them in a pull request from a stranger. - Keep a permission rule for anything that writes outside the project directory. The catalog is large and it grows every week, so the quality of any one skill is set by whoever published it. Reading the file is the whole review process, and it takes about the same time as installing it. ## How Tolz.me uses skills Tolz.me runs on skills. The repository holds one skill per job: writing and updating tool pages with their SEO metadata, generating FAQs and use cases, humanizing copy before it ships, and registering tools so browser agents can find them. Each one is a folder with a SKILL.md, checked in next to the code it maintains. The payoff is repeatability. A new tool page in the [Tolz.me tools directory](https://tolz.me/tools) gets the same metadata structure, the same FAQ shape, and the same pass over the writing, whether a person or a coding agent is doing the work. If you are setting up skills for the first time, a small utility with clear rules makes a good first target, since the right answer is easy to check. The [Cron expression translator](https://tolz.me/developer/cron-expression-translator) fits that description. For the other half of the problem, what happens after an agent writes the code, our guide on [shipping AI-written code you don't understand](https://tolz.me/blog/shipping-ai-written-code-you-dont-understand) covers the review side. ## What to take away A skill is a folder with a SKILL.md. Two required fields, a few optional ones, a name that has to match the folder, and a description that decides whether the agent ever opens the file. The agent sees the list, loads what it needs, and runs it with your permissions. That last part is the whole security story, and it is the reason to read the file before installing it. ## Sources - [GitHub weekly trending repositories](https://github.com/trending?since=weekly) - [OpenCode documentation: Skills](https://opencode.ai/docs/skills/) - [anthropics/claude-plugins-official](https://github.com/anthropics/claude-plugins-official) - [cursor/plugins](https://github.com/cursor/plugins) - [android/skills](https://github.com/android/skills) - [K-Dense-AI/scientific-agent-skills](https://github.com/K-Dense-AI/scientific-agent-skills) - [skills.sh skill catalog](https://skills.sh) ### chmod 755 vs 644 (and 600, 700, 777): Linux File Permissions Explained - URL: https://tolz.me/blog/chmod-755-vs-644-explained - Published: 2026-10-21 - Updated: 2026-10-21 - Tags: Linux, chmod, Permissions, Shell What the owner, group and others bits mean, how 644 and 755 are built from the r, w and x values, how directories differ from files, and which mode to pick. Two numbers decide who can read a file and who can run it. The owner, the group and everyone else each get one digit from 0 to 7, and 644 and 755 are just those digits written out. Once read is 4, write is 2 and execute is 1, the two modes stop being magic and start being arithmetic. ## Three classes, three permissions - Owner (u): the user who owns the file. - Group (g): the users who belong to the file's owning group. - Others (o): every other user on the system. - Read (r): open a file and see its contents. On a directory, list the names inside it. - Write (w): change a file's contents. On a directory, create, delete and rename the entries inside it. - Execute (x): run a file as a program. On a directory, enter it and reach the files inside. Wikipedia's write-up of file system permissions makes one point worth repeating: the effective permissions are the permissions of the class a user falls into first, in the order user, then group, then others. A user who owns a file uses the owner bits even when they are also in the owning group. ## Reading an ls -l listing The long listing prints the file type, then the three permission groups, then ownership and size. ``` -rwxr-xr-x 1 alice www-data 4096 Oct 21 09:14 deploy.sh ``` Take it in pieces: - The leading dash is the file type. A dash is a regular file, d is a directory, l is a symbolic link, c is a character device, b is a block device. - `rwx` is what the owner may do. - `r-x` is what the group may do. - `r-x` is what everyone else may do. - `alice` is the owner, `www-data` is the group, `4096` is the size in bytes, then the timestamp and the name. In this line the owner can read, change and run the file. The group and everyone else can read and run it but cannot change it. When a setuid, setgid or sticky bit is present, the third character of a group changes, so s or S takes the place of x in the owner and group columns, and t or T takes its place in the others column. ## The octal math: 4, 2, 1 Each permission is a single bit and each bit has a value. Read is 4, write is 2, execute is 1. The digit for a class is the sum of its bits. ``` 600 -rw------- owner can read and write 640 -rw-r----- the group can also read 644 -rw-r--r-- everyone can read 700 -rwx------ owner can do everything 755 -rwxr-xr-x everyone can read and traverse 777 -rwxrwxrwx everyone can do everything ``` Work through 644. The owner digit is 6, which is 4 plus 2, read plus write. The group digit is 4, read only. The others digit is 4, read only. So 644 means the owner can edit the file and everyone else can only read it. 755 uses 7 for the owner, which is 4 plus 2 plus 1, so read, write and execute. The group and others digits are 5, which is 4 plus 1, so read and execute. That is the standard mode for a program anyone can run and nobody but you can change. > Tip: The [chmod calculator](https://tolz.me/developer/chmod-calculator) converts between the octal form and the rwx letters, so you can build a mode from a description without doing the sum in your head. ## Files and directories store the same letters The bits are stored the same way for both, but the same letter does different work on a directory. - r on a directory lets you list the names of the files inside it. - w on a directory lets you create, delete and rename files inside it. - x on a directory lets you enter it and reach the files inside. The chmod(2) manual page calls this search permission: entries within the directory can be accessed. Two rules follow. Without x on a directory, w does nothing useful, because you still cannot reach the entries. With r but no x, you can see the names in the listing but opening them is a guessing game, as Wikipedia puts it. This is why web servers want 755 on directories and 644 on the files inside them. The server needs x to walk into the directory and r to read the files. A 644 directory is a dead end: the names are visible and nothing inside can be opened. ## Why 777 is a bad idea 777 gives read, write and execute to every account on the machine, including accounts that do not exist yet. On a shared host, any local user can read your files, change them and delete them. On a directory, any local user can drop a file into it. The dangerous part is what happens next. If a service account or a more privileged user later runs something from that directory, the code is already under someone else's control. There is one common exception, and it explains a mode you will see on every Linux box: ``` drwxrwxrwt 18 root root 4096 /tmp ``` /tmp is usually 1777, world writable with the sticky bit set. The sticky bit is what makes it safe. In a directory with the sticky bit, a user can only remove or rename files they own, which stops one user from deleting another user's uploads. World writable with a sticky bit is a deliberate pattern. World writable without one is 777. ## Symbolic chmod The octal form replaces the whole mode. The symbolic form changes specific bits, which is safer when you only want to add one thing. The format is who, then the operator, then the permissions. Who is u for the owner, g for the group, o for others, and a for all three. The operator is + to add, - to remove, or = to set. - `chmod u+x` adds execute for the owner only. - `chmod g-w` removes write for the group. - `chmod o=` clears every permission for others. The = operator sets the permissions you list and removes the ones you do not. - `chmod u=rw,go=r` sets the owner to read plus write and the group and others to read, which is 644. - `chmod a+x` adds execute for everyone. If you leave out the who, the change applies to all three classes and the umask decides which of them actually change. Writing out the class, as in u+x, is the habit worth keeping. ## umask: why new files are 644 and new directories are 755 The umask is a set of permissions that get turned off when a file or directory is created. The umask(2) manual page gives the typical default as 022, which is write for the group and write for others. Files are created with a request of 0666 and directories with 0777. The umask is subtracted from that request bit by bit: ``` files: 0666 & ~022 = 0644 directories: 0777 & ~022 = 0755 ``` That is why a text editor that saves a new file leaves it at 644, and a new directory lands at 755, without anyone typing chmod. If you want new files to stay private to your group, run umask 027 and the same file comes out as 640. Check your current value with the umask command. Since Linux 4.7, the umask of any running process is visible in the Umask field of /proc/pid/status. ## setuid, setgid and the sticky bit Three extra bits live in a fourth octal digit in front of the three permission digits. - setuid, 4000: a program with this bit runs with the privileges of the file's owner. The chmod(2) manual page describes it as setting the process effective user ID on exec. - setgid, 2000: on a directory, new files and directories created inside it inherit the directory's group. On an executable it sets the process group ID on exec. - sticky, 1000: on a directory, it stops users from removing or renaming files owned by other users even when they have write permission on the directory. That is the /tmp rule above. Set them with the leading digit: ``` chmod 4755 script # setuid chmod 2775 shared-dir # setgid on a directory chmod 1777 tmp # sticky bit ``` In ls -l output the setuid bit shows as s in the owner column, setgid as s in the group column and sticky as t in the others column. A capital letter means the execute bit is off in that position, so -rwsr-Sr-t is a setuid file that the owner can run and the group cannot. ## Modes for a web server A typical deployment uses a small set of modes. The web server needs to read files and walk through directories, and it rarely needs to change them. - Static files: 644. The server reads them and your deploy user can replace them. - Directories holding static files: 755, so the server can list and traverse them. - Secrets: 600 for the file and 700 for its directory. A config file with a database password or a private key should not be readable by every account on the box. - Upload or cache directories the application writes to: 775 with the group set to the application group, so the app user can write without giving write access to everyone. - Deploy scripts and admin tooling: 750 or 700, so only your group can run them. - Executables nobody should touch: 755 for the program, with the directory above it at 755 rather than 777. The pattern behind all of it: give every account the least it needs, keep write access off the directories a server reads from, and keep 777 away from anything that holds real data. ## Sources - GNU coreutils manual page, chmod(1): the symbolic and numeric mode formats, the u, g, o and a classes, the umask interaction and the sticky bit on directories. [man7.org](https://man7.org/linux/man-pages/man1/chmod.1.html) - Linux manual page, chmod(2): the permission bit values, the setuid, setgid and sticky bit definitions, and the execute or search permission on directories. [man7.org](https://man7.org/linux/man-pages/man2/chmod.2.html) - Linux manual page, umask(2): how the mask is applied to newly created files and directories, and the typical default of 022. [man7.org](https://man7.org/linux/man-pages/man2/umask.2.html) - File-system permissions, Wikipedia: the permission classes, the behaviour of directories, and the symbolic and octal notations. [en.wikipedia.org](https://en.wikipedia.org/wiki/File-system_permissions) ### Cron Syntax Cheat Sheet: 25 Real Schedules Explained - URL: https://tolz.me/blog/cron-syntax-cheat-sheet - Published: 2026-10-21 - Updated: 2026-10-21 - Tags: Cron, Linux, Scheduling, Shell The five fields of a cron expression, wildcards, lists, ranges, steps and names, the day-of-month or day-of-week rule, plus 25 schedules you can use. Cron runs commands on a schedule. The schedule lives in a crontab file, one job per line, and the time part of each line is a fixed sequence of five fields. Read them left to right and the rest is arithmetic. Read them in the wrong order and the job fires at hours you did not intend, or never fires at all. The crontab(5) manual page lists the fields in this order: minute, hour, day of month, month, day of week. Everything after the fifth field is the command to run. ## The five fields at a glance - Minute, 0 to 59. When inside the hour the job starts. - Hour, 0 to 23, in 24-hour time. 17 is 5 PM, 0 is midnight. - Day of month, 1 to 31. The calendar day number. - Month, 1 to 12, or a three letter name such as jan. - Day of week, 0 to 7, where 0 and 7 both mean Sunday, or a name such as mon. A job that runs every minute is written `* * * * *`, which reads as every minute, every hour, every day of the month, every month, every day of the week. A slightly more useful line looks like this: ``` 30 5 * * * /usr/local/bin/backup.sh ``` That is 05:30 every day. The command comes after the five fields, separated by whitespace. ## Wildcards, lists, ranges, steps and names - Wildcard `*` stands for every value in the field. The manual page describes it as first to last. - List: comma separated values. `1,15` means the 1st and the 15th of the month, `mon,wed,fri` means those three weekdays. - Range: two numbers joined by a hyphen. `9-17` covers 9 AM to 5 PM. Ranges are inclusive. - Step: a range or a wildcard followed by `/n`. `*/15` in the minute field fires at minutes 0, 15, 30 and 45 of every hour. - Names: the month and day of week fields accept the first three letters of the name, and case does not matter. `jan-mar` runs from January through March. Steps apply inside their own field only. That trips people up more than any other rule. `*/15` in the minute field is four runs an hour, not a run every fifteen minutes around the clock, because the minute field restarts at the top of every hour. The manual page makes the same point with `*/23` in the hours field, which means hour 0 and hour 23 of each day rather than every 23 hours. ## Sunday is both 0 and 7 The day of week field accepts 0 through 7, and the manual page states that 0 or 7 is Sunday. So `0 3 * * 0` and `0 3 * * 7` are the same schedule: 3 AM every Sunday. Using 0 lines up with the rest of the numeric fields, where counting starts at zero, and 1 is Monday. ## Short forms: @daily and friends The crontab(5) manual page lists a set of shortcuts that replace all five time fields. They are handy in scripts and self-documenting in a crontab file. - @reboot: run once after reboot. - @yearly and @annually: `0 0 1 1 *`, once a year on 1 January. - @monthly: `0 0 1 * *`, once a month on the 1st. - @weekly: `0 0 * * 0`, once a week on Sunday at midnight. - @daily: `0 0 * * *`, once a day at midnight. - @hourly: `0 * * * *`, once an hour on the hour. ## The day-of-month or day-of-week rule This is the rule that surprises people most. When both the day of month and the day of week fields are restricted, the job runs when either one matches. It does not wait for both. The manual page uses `30 4 1,15 * 5` as its example. That job runs at 4:30 AM on the 1st and 15th of every month, and it also runs at 4:30 AM on every Friday. The rule stays out of the way as long as one of the two fields is a wildcard. `0 12 * * 1` is every Monday at noon, because the day of month field is still `*` and only one field is restricted. > Tip: There is no syntax for the first Monday of the month. `0 9 * * 1` is every Monday, and `0 9 1 * 1` fires on the 1st of every month and on every Monday. Write the check in the script itself if you need that exact date. ## 25 schedules worth copying Every line below uses the standard five field format. Swap the command at the end for your own. - `* * * * *` runs every minute. - `*/2 * * * *` runs every other minute. - `*/5 * * * *` runs every five minutes. - `*/10 * * * *` runs every ten minutes. - `*/15 * * * *` runs at minutes 0, 15, 30 and 45 of every hour. - `0 * * * *` runs at the start of every hour. - `30 * * * *` runs at half past every hour. - `0 0 * * *` runs at midnight every day. - `0 12 * * *` runs at noon every day. - `0 9 * * *` runs at 9 AM every day. - `0 0,12 * * *` runs at midnight and noon every day. - `0 8-17 * * *` runs hourly from 8 AM to 5 PM. - `0 9-17 * * 1-5` runs every hour from 9 AM to 5 PM on weekdays. - `0 9 * * 1-5` runs at 9 AM on weekdays. - `0 9 * * 1,3,5` runs at 9 AM on Monday, Wednesday and Friday. - `0 9 * * mon` runs at 9 AM every Monday. - `30 6 * * 0` runs at 6:30 AM every Sunday. - `0 0 * * 0` runs at midnight on Sundays. - `0 22 * * 1-5` runs at 10 PM on weekdays. - `0 3 * * 1-5` runs at 3 AM on weekdays. - `0 0 1 * *` runs at midnight on the first day of every month. - `0 0 1 1 *` runs at midnight on 1 January every year. - `0 12 1,15 * *` runs at noon on the 1st and the 15th of every month. - `15 10 * * 6` runs at 10:15 AM every Saturday. - `*/10 * * * 1-5` runs every ten minutes on weekdays only. ## Time zones and daylight saving Cron matches each field against the clock on the machine running the daemon, so a schedule follows the server's local time. Move the server, change its time zone, or move the job to a different host, and the same expression now fires at a different wall clock time. The manual page spells out the daylight saving behaviour. Non-existent times, such as the missing hours during the conversion, never match, so a job scheduled during a missing hour does not run on that day. Times that occur twice, again during the conversion, cause matching jobs to run twice. In many time zones the change lands in the small hours, roughly between 1 AM and 3 AM. A job set for 2 AM can be skipped on the spring forward date and run twice on the fall back date. Schedules outside that window are far less surprising. - The manual page documents a CRON_TZ variable that sets the time zone for one crontab. Times in that table are read in that zone. - Log timestamps are written in the local time of the daemon, so a log line and a CRON_TZ schedule can look like they disagree by an hour or two. ## Testing a schedule before it goes live 1. Translate the expression before you install it. The [cron expression translator](https://tolz.me/developer/cron-expression-translator) turns a cron line into plain English and back, which catches a swapped field faster than reading five numbers left to right. 2. Check the syntax without installing the file. On systems running cronie, the `crontab -T` option tests a crontab file's syntax and stops at the first problem it finds. 3. Run the command by hand first. Cron uses a minimal environment: SHELL is set to /bin/sh, and HOME and LOGNAME come from the account's passwd entry. Use full paths for the command and for any file it touches. 4. Escape percent signs. The manual page notes that an unescaped `%` in a command turns into a newline, and everything after it is sent to the command as standard input. A date format such as `date +%d` needs a backslash. 5. Send the output somewhere you will read it. Cron mails output to the crontab owner by default, so set MAILTO to an address you check or redirect to a log file with `>> /var/log/myscript.log 2>&1`. 6. Give it a minute to start. The daemon examines crontab entries every minute, so a new job may not run until the next minute boundary. A Unix timestamp pins an exact instant. When you want to confirm a schedule lands on the date you expect, take the next run time, convert it to a timestamp with the [Unix timestamp converter](https://tolz.me/developer/unix-timestamp-converter), and read it back in your own time zone. ## Mistakes that show up again and again - Field order. The minute comes first and the month comes fourth. `0 6 1 * *` is 6 AM on the 1st of every month, while `0 6 1 1 *` is 6 AM on 1 January. - Steps reset with their field. `*/15` in the minute field is four runs an hour, not a run every fifteen minutes. - Sunday numbering. 0 and 7 are both Sunday and 1 is Monday, which does not match the calendar order most people carry in their head. - Missing fields. Cron expects five of them. Four is a syntax error, not an implied midnight. - Forgetting the OR rule. `0 9 1 * 1` is not the first Monday of the month. It runs on the 1st of every month and on every Monday. Most of these disappear once the expression has been read back to you in words. Translate it, check the timezone, then install it. ## Sources - Linux manual page, crontab(5): the field ranges, lists, ranges, steps and names, the day-of-month or day-of-week rule, the `@` shortcuts, the CRON_TZ variable and the daylight saving behaviour. [man7.org](https://man7.org/linux/man-pages/man5/crontab.5.html) - Linux manual page, crontab(1): the `-l`, `-e` and `-T` options. [man7.org](https://man7.org/linux/man-pages/man1/crontab.1.html) - Cron, Wikipedia: background on cron and the wider family of cron-like schedulers. [en.wikipedia.org](https://en.wikipedia.org/wiki/Cron) ### Discord Text Formatting Guide: Every Style With Examples - URL: https://tolz.me/blog/discord-text-formatting-guide - Published: 2026-10-21 - Updated: 2026-10-21 - Tags: Discord, Formatting, Markdown A practical Discord markdown reference for bold, headers, subtext, masked links, code blocks, ANSI colors, spoilers, timestamps, and small text. Discord formatting is markdown with a few platform-specific extras. Most of it is quick to type once you have used it a few times, but the edge cases matter. A header that starts one space too far right stays a header only if the client agrees, a masked link works in one place and not another, and colored code blocks look different on a phone. This reference covers the styles you will actually use, with an example for each. Keep the [text formatter](https://tolz.me/discord/text-formatter) open if you would rather copy the syntax than type it. ## Bold, italic, underline, and strikethrough ``` **bold** *italic* __underline__ ~~strikethrough~~ ***bold italic*** ``` Bold uses two asterisks on each side. Italic uses one. Underline uses two underscores. Strikethrough uses two tildes. You can combine bold and italic with three asterisks, and some mixed combinations such as bold plus underline also parse. Nested styles get less predictable across clients, so check the preview if the formatting carries meaning. Do not use underscores for emphasis inside a word. Discord may read them as literal characters or create unexpected italics. Asterisks are safer for general emphasis. Save bold for labels and short warnings, not entire paragraphs. A wall of bold text is harder to read than plain text. ## Headers ``` # Large header ## Medium header ### Small header ``` A header marker has to be at the start of a line, followed by a space. Discord renders three heading sizes. They work well in pinned messages, server guides, and long announcements where members need to scan sections. They are less useful in a fast-moving chat, where a large heading pushes the conversation off screen. If a header does not render, check for a leading space before the `#`. That one space is enough to make Discord treat the line as ordinary text. ## Subtext ``` -# This line renders as subtext ``` Subtext is the small gray line you see under some Discord announcements. The `-#` prefix must start the line and be followed by a space. It is a message format, not a font, so it will not work in a display name or nickname. Every line needs its own prefix. Use it for timestamps, footnotes, and secondary details. Do not put rules or warnings in subtext. Small gray text is easy to skip, especially on a phone. ## Lists ``` - First item - Second item - Nested item 1. First step 2. Second step ``` Discord accepts hyphens or asterisks for bullet lists and numbers followed by a period for ordered lists. Indent nested items with spaces. Keep each item short. A list with long wrapped sentences stops looking like a list and starts looking like paragraphs with dots. For a checklist in a pinned message, a plain bullet list is usually clearer than a row of emoji. If you need actual checkboxes, a bot or an external task tool is a better fit. Discord does not turn `[ ]` into an interactive checkbox. ## Block quotes ``` > This line is quoted. >>> This line and everything after it is part of the same block quote. ``` A single `>` quotes one line. Three of them (`>>>`) quote the rest of the message, which is handy for pasting a longer excerpt without adding a marker to every line. Quotes are useful for rules, support replies, and separating an announcement from the discussion around it. ## Inline code and code blocks ``` Use `npm run build` to check the project. ```js const message = "Hello"; console.log(message); ``` ``` Single backticks style a short piece of text as inline code. That is the right choice for file names, commands, and short strings. Triple backticks create a code block, which keeps spacing and turns off markdown inside the block. Adding a language tag after the opening fence enables syntax highlighting for that language. `js`, `json`, `python`, `css`, `html`, and many others are recognized. Highlighting depends on the client, so a block that looks colored on desktop may appear plain elsewhere. The text itself is unchanged. Code blocks are also the safest way to show markdown syntax to someone else. A code block full of asterisks will not turn into bold text. ## ANSI colored code blocks ``` ```ansi \u001b[0;31mRed text\u001b[0m \u001b[0;32mGreen text\u001b[0m \u001b[0;34mBlue text\u001b[0m ``` ``` The backslash-u sequence above stands for the escape character that starts each ANSI sequence. Discord needs the real character, not the printed notation, which is why copying a block from a generator is safer than retyping it. Discord can render a limited set of ANSI colors inside a code block tagged `ansi`. The block contains escape sequences around the text. The [color text generator](https://tolz.me/discord/color-text-generator) builds those sequences for you so you do not have to paste invisible control characters by hand. Colored text is a desktop feature in practice. Some mobile clients show the code block in monospace without applying the colors. Never use color alone to carry meaning. If red means urgent, add the word urgent as well. ## Masked links ``` [Tolz.me tools](https://tolz.me/tools) ``` A masked link shows your label instead of the full URL. The syntax is regular markdown: square brackets for the label, parentheses for the address. The important catch is where it works. As of writing, masked links render in embed descriptions and field values, and in embeds sent by bots or webhooks. They do not reliably render in ordinary user messages, where the raw markdown may show instead. If you need a clickable label in a normal message, send the plain URL. If you control the embed, use the markdown link and test it in a private channel first. ## Spoilers ``` The ending is ||the ship sinks||. ||A whole hidden sentence.|| ``` Double pipes hide text behind a gray bar until someone clicks or taps. You can spoil a single word, a whole line, or several adjacent sections. Nested spoilers are not reliable, so keep one layer. A pipe character inside the hidden text can end the spoiler early, which is why the [spoiler text generator](https://tolz.me/discord/spoiler-text-generator) checks for stray pipes. Images use a different mechanism. Attach the image, then choose Mark as spoiler in the upload preview before sending. On mobile, select the attachment and use the spoiler option in the preview. Bot uploads can prefix the filename with `SPOILER_`. ## Timestamps ``` ``` Timestamps use a Unix time in seconds inside a `` tag. The letter after the second colon sets the style. `t` is short time, `T` is long time, `d` is short date, `D` is long date, `f` is short date and time, `F` is long date and time, and `R` is relative time such as in two hours. The value is always the same moment. Discord translates it into each reader's local timezone, which is the reason to use a timestamp instead of typing a time by hand. The [timestamp generator](https://tolz.me/discord/timestamp-generator) converts a date and lets you compare styles before copying. ## Unicode small text and fancy names There are two different kinds of small text. Subtext uses the `-#` prefix and works in messages. Unicode small caps, superscript, and subscript use separate characters that look smaller or styled. They work in display names, nicknames, bios, and messages, but they are not real font changes. Each device renders them with the fonts it has installed. Unicode styles have gaps. Not every letter has a subscript form, and screen readers may pronounce small caps letter by letter or skip them. Use them for short labels and decorative names, not for instructions or anything a member must understand on the first read. The [small text generator](https://tolz.me/discord/small-text-generator) shows which letters convert cleanly. Discord usernames are stricter than display names. They allow lowercase letters, numbers, periods, and underscores, so fancy Unicode usually fails there. Put styled text in your display name or server nickname instead. ## Combining styles without making a mess - Use one heading level for the main sections of a long post and subtext for the small details. - Bold a label, then write the value in plain text. Do not bold the value too. - Use inline code for commands and exact text, and code blocks for anything that depends on spacing. - Keep ANSI color for short highlights. It is not a substitute for structure. - Use spoilers for genuine surprises and content warnings, not for ordinary text that people need to read. A useful test is to strip the formatting and read the message as plain text. If it still makes sense, the formatting is supporting the content. If it falls apart, the structure was doing too much work. ## Client differences to expect Discord desktop, web, iOS, and Android do not render every style the same way. Headers and code blocks are broadly consistent. ANSI colors are the least consistent. Spoiler reveal gestures vary. Masked links depend on whether you are inside an embed. When a post matters, check it on the device your members use most. Formatting is also only half the message. A well-formatted announcement still needs a clear time, a short explanation, and one obvious next step. The style should make that easier to find, not compete with it. ### How to Calculate Hours Worked (With a Lunch Break and Overnight Shifts) - URL: https://tolz.me/blog/how-to-calculate-hours-worked-with-lunch-break - Published: 2026-10-21 - Updated: 2026-10-21 - Tags: Hours, Timesheet, Shift Work, Calculators A step-by-step way to total a shift: subtract an unpaid lunch, handle shifts that cross midnight, and convert minutes to decimal hours for a timesheet. Most timesheet errors come from one of three places: a lunch break that was never subtracted, a shift that crosses midnight, and minutes converted to decimal in a hurry. All three are easy to get right once you work in minutes instead of hours. ## The method 1. Write the start and end time in 24-hour format. 5:30 PM becomes 17:30. 2. Convert both to minutes since midnight. 9:00 is 540, 17:30 is 1050. 3. Subtract. If the end time is smaller than the start time, the shift crosses midnight, so add 1440 minutes to the end time first. 4. Subtract the unpaid break. A 30-minute lunch comes out of the total. 5. Divide the minutes by 60 to get decimal hours. 480 minutes becomes 8.00. Stay in minutes the whole way. Rounding the hours at the start is where mistakes get baked in. The 1440 in step 3 is the number of minutes in a day. Once a shift passes midnight, the end time is a smaller number than the start time, so a day has to be added back before the subtraction makes sense. ## Example 1: 9:00 to 17:30 with a 30-minute lunch ``` start 9:00 = 540 minutes end 17:30 = 1050 minutes span = 1050 - 540 = 510 minutes lunch = -30 minutes total = 480 minutes decimal = 480 / 60 = 8.00 hours ``` The shift covers eight and a half hours of clock time. Take out the 30-minute unpaid lunch and the answer is 8.00 hours. ## Example 2: 8:45 to 17:15 with a 45-minute lunch ``` start 8:45 = 525 minutes end 17:15 = 1035 minutes span = 1035 - 525 = 510 minutes lunch = -45 minutes total = 465 minutes decimal = 465 / 60 = 7.75 hours ``` 465 divided by 60 is exactly 7.75, so nothing needs rounding. This is the same span of clock time as example 1 with a different lunch, and that is where the half hour difference comes from. ## Example 3: 22:00 to 06:30 with a 30-minute break ``` start 22:00 = 1320 minutes end 06:30 = 390 minutes end crosses midnight, so 390 + 1440 = 1830 span = 1830 - 1320 = 510 minutes break = -30 minutes total = 480 minutes decimal = 480 / 60 = 8.00 hours ``` The rule to remember: when the end time is earlier than the start time, add a full day, 1440 minutes, to the end time before subtracting. A shift like this also spans two calendar dates, so a timesheet that records the date as well as the time will show the shift on both days. ## Minutes to decimal hours Divide by 60. These are the values that come up most often. - 5 minutes = 0.08 hours - 10 minutes = 0.17 hours - 15 minutes = 0.25 hours - 20 minutes = 0.33 hours - 30 minutes = 0.50 hours - 45 minutes = 0.75 hours - 60 minutes = 1.00 hour - 90 minutes = 1.50 hours - 120 minutes = 2.00 hours The two values that look awkward, 0.08 and 0.17, are rounded. 5 divided by 60 is 0.0833 and 10 divided by 60 is 0.1667. If the payroll system wants exact minutes, submit the minutes and let the system do the division. ## Why timesheets use decimal hours Decimal hours make a pay period easy to total and easy to pay. 7.75 hours at 22.00 an hour is 170.50, which is a much easier sheet to check than 7 hours and 45 minutes multiplied by a rate. Keep the minutes as your working number, then convert to decimal when you write the total. ## Five totals worth memorising - 8:30 to 17:00 with a 30-minute lunch: 510 minutes of clock time, 480 after the break, 8.00 hours. - 7:00 to 17:00 with a 30-minute lunch: 600 minutes, 570 after the break, 9.50 hours. - 12:00 to 20:30 with a 30-minute lunch: 510 minutes, 480 after the break, 8.00 hours. - 22:00 to 06:00 with a 30-minute break: 480 minutes across midnight, 450 after the break, 7.50 hours. - 9:00 to 13:00 and 14:00 to 18:00: 240 minutes plus 240 minutes, 8.00 hours, with the unpaid hour in the middle already excluded. ## Weekly totals Add the minutes for the week first, then convert once. Converting each day separately and adding the decimals gives the same answer, but only when nothing was rounded on the way. ``` Monday 480 minutes Tuesday 465 minutes Wednesday 480 minutes Thursday 510 minutes Friday 360 minutes week total = 2295 minutes decimal = 2295 / 60 = 38.25 hours ``` Thirty-eight hours and fifteen minutes. If that week had one fewer 15-minute break, the total would be 2280 minutes, which is exactly 38.00 hours. ## Which breaks come out of the total Whether a break is deducted is a policy question, not an arithmetic one. A meal break is often unpaid because you are free to leave the workplace, while a short rest break is often treated as worked time. Wikipedia's overview of working time notes that the hours people actually work vary in practice precisely because some employers count breaks as part of the working week and others do not. The same shift can therefore be 8.00 or 8.50 hours depending on the rule. Settle that first, then do the math. ## Rounding: check the rule before you round Rounding up looks generous and increases the payroll total. Rounding down does the opposite. Some employers round to the nearest quarter hour, some to the nearest 15 minutes, and some do not round at all. Watch for rounding that works against the worker. Wikipedia's write-up of California overtime law gives a specific example: rounding a six-hour day and a ten-hour day into two eight-hour days would take two hours of overtime pay away from the worker. When your employer rounds, the direction matters. A habit that holds up well is to keep the raw minutes in the sheet and round once, at the end of the week or the pay period, following the employer's stated rule. ## Edge cases worth writing down - A shift that crosses midnight: add 1440 minutes to the end time, as in example 3. - A break that crosses midnight, for example 23:45 to 00:15. Convert both ends the same way and the break is 30 minutes. - Two shifts in one day, for example a split shift. Calculate each one separately and add the two totals. - A 24-hour shift: 1440 minutes, which is 24.00 hours. Check how your employer records a full day before you submit it. - Times written in 12-hour format: convert them first, so 5:30 PM does not quietly become 5:30 AM. ## Let the calculator do the subtraction The [hours between times calculator](https://tolz.me/calculator/hours-between-times) takes two clock times and gives you the span between them, so you can subtract the break and read the decimal straight away. It is a useful second check when you are filling in a timesheet by hand. ## One last check on the rules This is arithmetic, not legal advice. Whether a break is unpaid, how rounding works and when overtime starts come from your employment contract, your employer's policy and the labour law where you work. Wikipedia's summary of overtime rules shows how much these differ between places: the European Union working time directive caps the average working week, including overtime, at 48 hours over a 17-week reference period, asks for a break when the working day runs past six hours, and requires at least 11 consecutive hours of daily rest in every 24. United States federal law sets a 40-hour work week with overtime paid at one and a half times the normal rate for many workers. Check your own rules before you file the numbers. > Tip: Keep the minutes, convert once, and write down which break rule you used next to the total. That single line saves an argument later. ## Sources - Working time, Wikipedia: why the hours people actually work vary between employers, and how standard working weeks differ by country. [en.wikipedia.org](https://en.wikipedia.org/wiki/Working_time) - Overtime, Wikipedia: the European Union working time limits, United States federal overtime rules, and the California example about rounding workdays. [en.wikipedia.org](https://en.wikipedia.org/wiki/Overtime) ### How to Make a Table in Discord (That Actually Lines Up) - URL: https://tolz.me/blog/how-to-make-a-table-in-discord - Published: 2026-10-21 - Updated: 2026-10-21 - Tags: Discord, Formatting, Tables Discord does not render markdown pipe tables. Here is how to build a clean monospace table in a code block, with worked examples and fixes for mobile. You paste a neat markdown table into Discord and it lands as a row of pipes and dashes. No columns. No rows. Just text. Discord does not render pipe tables, and there is no table button hiding in the message box. The workaround is a text table inside a code block, and it looks good once you know which parts to control. ## Why pipe tables do not render Discord supports a small set of markdown: bold, italic, underline, strikethrough, headers, subtext, spoilers, inline code, code blocks, block quotes, and lists. Tables are not on that list. A line like `| Player | Score |` is treated as ordinary characters, so the pipes show up literally. Even if Discord did read the pipes, normal message text uses a proportional font. An `i` is narrower than an `m`, so a column built from spaces drifts as soon as the content changes. A code block switches to a monospace font where every character takes the same width. That is the whole trick. The table is still plain text, but the spacing holds. Use the [Discord table generator](https://tolz.me/discord/table-generator) to build the grid without counting spaces by hand. It handles the borders, padding, and a character counter so you can see whether the message will fit before you post. ## Build a table step by step 1. Open the [table generator](https://tolz.me/discord/table-generator) and set the number of rows and columns. Keep the header row on if the first row is a label row. 2. Type into the cells, or paste comma-separated or tab-separated data from a spreadsheet and choose the fill option. Tab-separated data from Excel or Google Sheets is the most reliable. 3. Pick a border style. Box drawing looks like a drawn table, ASCII is safer across fonts, and minimal drops the frame for short lists. 4. Set alignment per column. Left for names and labels, right for numbers, center only for short markers. 5. Copy the full output, including the opening and closing triple backticks. Those fences are what make the table monospace. 6. Paste it into Discord and check the result on a phone before you pin or announce it. ## What the code block looks like A minimal table is just three lines of text wrapped in fences. The fences need to sit on their own lines: ``` ``` ┌──────────┬───────┐ │ Session │ Time │ ├──────────┼───────┤ │ Opening │ 18:00 │ │ Workshop │ 19:30 │ └──────────┴───────┘ ``` ``` You can add a language tag after the first fence, but it does not change the table. It only tells Discord to attempt syntax highlighting. For a plain table, leave it off. Two things break the block. A missing closing fence leaves the rest of your message inside the code block. A cell that contains three backticks closes the fence early, so keep backticks out of cell text or replace them with a different character. ## Worked example: a raid schedule Suppose your server runs three sessions on a raid night. Copy this into the generator's import box: ``` Session,Time,Host Opening,18:00,Ari Boss run,19:30,Bea Cleanup,21:00,Cyrus ``` With box borders, a header row, and left alignment for every column, the copied message looks like this: ``` ┌──────────┬───────┬───────┐ │ Session │ Time │ Host │ ├──────────┼───────┼───────┤ │ Opening │ 18:00 │ Ari │ │ Boss run │ 19:30 │ Bea │ │ Cleanup │ 21:00 │ Cyrus │ └──────────┴───────┴───────┘ ``` The times are short enough that left alignment reads fine. If you add time zones or longer labels, right alignment for the time column keeps the digits together and makes the schedule easier to scan. ## Worked example: a leaderboard Numbers are where alignment earns its keep. A leaderboard with left-aligned points looks ragged because 9 and 900 start at the same place. Right alignment puts the last digit in a straight line: ``` ┌──────────┬────────┬──────┐ │ Player │ Points │ Wins │ ├──────────┼────────┼──────┤ │ Nova │ 2450 │ 12 │ │ Kite │ 980 │ 5 │ │ Ember │ 120 │ 1 │ └──────────┴────────┴──────┘ ``` That table is a snapshot. Discord does not update it when someone plays another match. If your community needs a live scoreboard, use a bot command or a linked spreadsheet. A code block table is for information that is already final when you post it. ## Pick the right border style - Box drawing uses characters like `┌`, `─`, and `│`. It looks like a real table and suits announcements, schedules, and pinned reference posts. - ASCII uses `+`, `-`, and `|`. It is less pretty, but it survives fonts, terminals, and copy-paste into other apps with fewer surprises. - Minimal removes the frame and separates columns with spaces. It works for two or three short columns and feels lighter than a full grid. Border characters are also characters. A wide table with long borders costs more of the message limit than the same content with minimal borders. That rarely matters until you build something with many columns. ## Emoji, CJK text, and alignment problems Monospace does not mean every character is one cell wide. Chinese, Japanese, and Korean characters usually take two cells. Most emoji take two, and some emoji are built from several code points joined together, so a single visible symbol can take more space than expected. A good generator measures display width rather than counting characters, which gets common CJK text and emoji closer to the correct padding. It still cannot account for every font or every Discord client. If a column drifts after you post, the emoji is usually the cause. Move it to its own column or remove it. Combining marks, skin tone modifiers, and zero-width joiners make the math harder. Keep cells short, avoid decorative emoji inside tight columns, and use plain letters and numbers when exact alignment matters more than personality. The [text formatter](https://tolz.me/discord/text-formatter) is useful for styling the heading or note around the table, but keep the table itself plain. ## Why the table wraps on mobile Mobile clients often wrap long code block lines instead of scrolling sideways. The spaces are still there, and copying the message preserves them, but the visual grid breaks across extra lines on a narrow screen. Plan for the smallest screen your members use. Three or four short columns usually fit. Eight columns may look fine on a desktop monitor and turn into a wall of broken lines on a phone. Shorten headers, remove optional columns, or split a wide table into two narrower posts. ## When embed fields are the better choice Bots and webhooks can send embeds with up to 25 fields, and each field has a name and a value. Two or three inline fields can imitate a small table with a cleaner layout and no monospace spacing to maintain. The catch is that fields are not a real grid either. Long values wrap inside their column, and field names cannot carry every kind of markdown. Use embed fields when a bot is already sending the message and the data fits into short label and value pairs. Use a code block table when a person is posting in chat and alignment is the point. The [embed generator](https://tolz.me/discord/embed-generator) helps you test the field layout before handing it to a bot or webhook. ## The 2,000 and 4,000 character limits At the time of writing, a standard Discord message holds 2,000 characters, and Nitro raises that to 4,000. Every space used for padding, every border character, and both code fences count toward the total. A table that fits under 2,000 is safe for everyone. If it only fits under 4,000, some members may not be able to send or copy the same message. Split a long table into two messages and keep a complete set of fences around each half. Do not leave one code block open across two messages. ## Troubleshooting - Visible backticks mean the code fence is broken. Check that the opening and closing fences each sit on their own line. - Uneven columns usually mean part of the table is outside the code block. Select the whole block and paste it again from the tool. - One column of jumbled text means the separator was not recognized. Copy with tabs from a spreadsheet, or use commas and quote fields that contain commas. - A rejected message is often over the limit. Remove an empty column, shorten long labels, or switch to minimal borders. - A grid that breaks on a phone needs fewer columns, not smaller text. Code blocks do not get a horizontal scrollbar on every mobile client. The final check is simple. Read the table in the Discord preview, then read it on a phone. If the columns hold and the message fits, you are done. If not, change the layout rather than fighting the spacing. ### MiniMax M3.1 Flash: What's New, What It Costs, and What the Benchmarks Do and Don't Tell You - URL: https://tolz.me/blog/minimax-m3-1-flash-what-is-new - Published: 2026-10-21 - Updated: 2026-10-21 - Tags: AI Models, MiniMax, LLM, Cost MiniMax claims faster inference and a 70% lower agent cost for M3.1 Flash. Here is the price math and how to test a model on your own work. MiniMax announced M3.1 Flash on October 19, 2026. The announcement leads with speed and price: nearly three times faster online inference, and 70% lower usage cost for agentic tasks than the previous model. The release drew a busy Hacker News discussion. ## What MiniMax says changed - An early version of the model ran anonymously under the name "Space Bunny". MiniMax says that version processed more than 200 trillion tokens. - Online inference is nearly 3x faster than the predecessor. - Usage cost on agentic tasks is 70% lower. - Output speed runs at 200+ tokens per second, with a median of 170+. - The model is available through the MiniMax API, the M Plan, and MiniMax Code. - Weights will be released later, and a Pro model is planned for later still. That last pair deserves attention. Open weights and a Pro tier are both on the roadmap rather than available today. If you need to self-host now, the wait is part of the product, not an afterthought. ## The price list Per one million tokens, as published by MiniMax: - Input: $0.15 - Output: $0.60 - Cached input: $0.015 Output costs four times input, which is normal for the industry and worth noticing if your workload generates a lot of tokens. The cached input rate is a tenth of the input rate. Both numbers matter, and the second one matters more than most people expect. ## The cost math on one agent run Here is a concrete shape: an agent session that sends 2 million input tokens and returns 500 thousand output tokens. ``` Input: 2,000,000 tokens x $0.15 / 1M = $0.30 Output: 500,000 tokens x $0.60 / 1M = $0.30 Total: $0.60 ``` Sixty cents for a session that pushed 2.5 million tokens through it. If the 2 million input tokens are served from cache instead, the same session costs: ``` Cached input: 2,000,000 tokens x $0.015 / 1M = $0.03 Output: 500,000 tokens x $0.60 / 1M = $0.30 Total: $0.33 ``` Now scale it up. A thousand sessions like that in a month comes to $600, or $330 if the input is cached. If your agent loops and re-sends the same context on every turn, caching is the difference between a bill you notice and a bill you do not. ## Cost per finished task is the number that counts The 60 cents is per session, not per result. An agent that needs five attempts to finish one task has spent $3.00 on that task, and the gap between two models on price is often smaller than the gap in attempt count. This is why cost per finished task is the metric to track. A model at half the price that takes twice as many attempts is the expensive one. ## What the benchmarks do and don't tell you MiniMax published five scores with the announcement. The short labels below say what each benchmark name refers to. The scores are MiniMax's: - BankerToolBench: 76.2, a tool-use result built around finance tasks. - DeepSWE v1.1: 73.7%, a software-engineering agent result. - OSWorld V2: 61.7, an operating system and computer-use result. - Agents' Last Exam (ALE): 29.6, which MiniMax describes as testing real professional work and complex tasks, and the lowest score on the list. - CyberGym: 82%, a security result. Two things to keep straight about that list. First, the numbers are vendor-reported. Second, a score does not carry its own methodology. You cannot rerun these from a blog post, because the harness, the prompts, and the scoring rules live with whoever ran them. MiniMax also says the model ranks first among Chinese models on OSWorld V2 and on ALE. That is a narrower claim than topping those benchmarks overall, and it comes from the same source as the scores. What the list is good for is showing where to look. CyberGym at 82% and BankerToolBench at 76.2 sit next to ALE at 29.6. A model can do well on a targeted benchmark and still be far from dependable on open-ended agent work, where the goal is ambiguous and the task runs for many steps. If that is your use case, the low score on the hardest evaluation is the number to read twice. ## Why two vendors' scores rarely line up Scores from different announcements are rarely comparable, even when they name the same benchmark. Each lab runs the suite inside its own harness, with its own agent scaffolding, its own retry policy, and its own version of the task set. A model that scores higher on one leaderboard and lower on another has not necessarily changed. The setup changed. The practical consequence is that cross-vendor comparisons are close to meaningless, and even the same vendor's numbers across releases are only comparable when the harness is unchanged. Treat each release note as a set of claims to be checked, not a ranking to be trusted. ## How to check the speed and cost claims yourself The three headline claims are easy to test on your own traffic: - Speed: run your longest real prompts, not short greetings. A model that reports 200+ tokens per second on a 200-token reply tells you nothing about a 4,000-token one. - Cost: log your cache hit rate before switching. The $0.015 cached input rate only turns into savings if your workload actually reuses context. Check how often yours does. - Agent cost: compare tokens spent per finished task against the model you use now. If the 70% lower usage cost claim holds on your prompts, this is where it shows up. One more thing to confirm before you move a workload: what access through the MiniMax API, the M Plan, and MiniMax Code actually gives you in terms of rate limits and concurrency. A cheap token price is not useful if the endpoint throttles the agent loop that needs it. ## How to test a model on your own work A vendor leaderboard cannot tell you whether the model handles your codebase, your ticket format, or your API. Build a small test set instead. 1. Collect twenty tasks from your own backlog. Real tickets, real bugs, real requests. 2. Write the prompt you would actually give the agent, not a friendly demo prompt. 3. Per task, record whether it finished, how many attempts it took, how long it took, and how many tokens it used. 4. Run the same set on the model you use today. You need the comparison, not the absolute number. 5. Multiply the cost of a single run by the number of attempts. That is your real unit cost. Small utilities like the [Unix timestamp converter](https://tolz.me/developer/unix-timestamp-converter) make good test tasks, because the right answer is easy to check and the edge cases are known. That is a good shape for a test set: a clear correct answer, a few cases that break naive solutions, and no ambiguity about what finished means. ### Latency MiniMax reports output at 200+ tokens per second with a median of 170+. For an agent, latency shows up as time per turn, and time per turn compounds across a long task. At 170 tokens per second, a 500-token answer takes roughly three seconds. At 200, roughly two and a half. ### Tool calling The tool-use scores are the part most likely to transfer to your work, and also the part least like your work. Real tool calling breaks on things a benchmark does not cover: your tool names, your parameter shapes, your error messages, your rate limits. Test the exact tools the agent will call, with the exact schemas it will see. ### Reliability over peak performance Check that the model handles a tool that returns an error, a tool that returns an empty result, and a task where the right move is to stop and ask a person. Those three cases show up constantly in real agent runs, and they separate models more often than the headline numbers do. For reference, an independent model tracking page exists at [Artificial Analysis](https://artificialanalysis.ai/models/minimax-m3-1-flash). It is worth a look if you want numbers from a third party rather than from the vendor. We do not repeat its figures here. ## What to take away MiniMax M3.1 Flash is a fast, cheap model with open weights still to come. The vendor numbers look encouraging on tool use and security tasks, and the pricing is simple enough to reason about without a spreadsheet. The cheapest way to find out whether it fits your work is to run twenty of your own tasks through it and compare cost per finished task against whatever you use today. Tolz.me builds with AI coding agents and tests models on real tasks rather than published leaderboards. It is the evaluation that actually decides the question for your code, and it is the one most teams skip. ## Sources - [MiniMax announcement: M3.1 Flash, October 19, 2026](https://www.minimax.io/blog/minimax-m31f) - [Hacker News discussion](https://news.ycombinator.com/item?id=50154540) - [Artificial Analysis model page](https://artificialanalysis.ai/models/minimax-m3-1-flash) ### Roblox Display Name Rules: What Works and What Gets Rejected - URL: https://tolz.me/blog/roblox-display-name-rules - Published: 2026-10-21 - Updated: 2026-10-21 - Tags: Roblox, Gaming, Names The difference between a Roblox username and display name, the published character rules, the 7-day change limit, and fixes for common rejections. Roblox gives every account two names: a username and a display name. They look similar in a game, but they follow different rules and serve different purposes. Most rejected display name attempts come from people applying username habits to the display name field. Here is how the two names differ, what the published display name rules allow, and why a name that looks clean can still fail when you press save. ## Username vs display name Your username is the unique handle attached to your account. You use it to log in, friends use it to find you, and nobody else can have the same one. Usernames follow a stricter set of rules: letters and numbers, a limited length, and at most one underscore. They are not meant to be changed often. Your display name is the label other players see above your avatar, in chat, and in friend lists. It is not unique. Two accounts can use the same display name at the same time. That is why there is no availability check for display names, and why copying a well-known name does not prove anything. Display names can include spaces and underscores, which makes them friendlier for names like Moon Rider or Luna_7. They still have limits, and the text filter still applies. ## The published display name rules Roblox can change its rules, and its support pages are the authoritative source. The checks below reflect the rules published at the time of writing. - Length: 3 to 20 characters. Spaces count toward the total, so a two-word name uses more of the limit than it looks like. - Characters: letters, numbers, spaces, and underscores. Symbols such as stars, brackets, hearts, and punctuation marks are outside that set. - Spacing: start and end with a visible character. A leading or trailing space is easy to create when pasting from another app. - Filter: the name must pass Roblox's text filter, which checks for names that break the community rules. - Changes: Roblox allows a display name change once every seven days on the same account. These are not the username rules. A username has a different length range and does not accept spaces. If you are checking a login handle, use the username guidance instead of a display name tool. ## Why symbols and emoji get rejected The display name field accepts a narrow character set. Emoji, decorative stars, hearts, arrows, and punctuation are not letters, numbers, spaces, or underscores, so they fail the character check before the name ever reaches the moderation filter. Fancy Unicode fonts are a common trap. A copied aesthetic font may look like ordinary letters on screen, but the characters are different code points. To Roblox they are not letters, so the field rejects them. Accented letters such as é or ü are usually rejected for the same reason, even though they are normal in many languages. Zero-width characters and lookalike letters cause another kind of problem. They may sneak past a simple character count while making the name unreadable or easy to confuse with someone else's. Using them to get around the filter is a moderation risk, not a clever workaround. If you want a symbol in your identity, put it somewhere that accepts it. A game profile, a group description, or a Discord nickname has different rules. The display name itself is not the place for decoration. ## Why a name can pass the character check and still fail Roblox runs a text filter on display names. The filter looks for slurs, sexual content, harassment, personal information, and attempts to impersonate Roblox staff or popular creators. It also catches some innocent strings that happen to contain a blocked sequence across a word boundary. That is why two ordinary words can fail when they are joined. The filter does not know your intent, and no outside checker can reproduce it exactly. If Roblox rejects a name that passes every local check, change the wording instead of trying to disguise it. Names that include a full name, school, age, location, or contact details can also cause problems. A display name is public. Keep personal information out of it, even if the field accepts the characters. ## How to change your display name on PC 1. Sign in to Roblox and open the account menu. 2. Open Settings, then find Account Info. 3. Select the display name field and choose the edit option. 4. Type the new name and confirm the change. 5. If the save fails, read the error message before trying again. It usually points to length, characters, or the filter. Menu labels move between Roblox versions, so the exact wording may differ. Look for Account Info and Display Name rather than memorizing a single path. ## How to change it on mobile 1. Open the Roblox app and sign in. 2. Tap the More or three-dot menu, then open Settings. 3. Go to Account Info. 4. Tap Display Name, enter the new name, and save. If the option is missing, update the app first. Some account settings are easier to find in a browser, so you can also sign in on the Roblox website from a phone and use the desktop layout. ## The seven-day change limit Roblox allows a display name change once every seven days. The timer belongs to the account, not to a device or a game. If you changed your name yesterday, you need to wait before the next one goes through. That limit makes it worth testing a name before you commit. Use the [display name checker](https://tolz.me/gaming/roblox-display-name-checker) to catch length, character, and spacing problems first. It cannot run the Roblox filter or tell you when your account last changed its name, but it removes the obvious mistakes. Roblox's support page is the place to confirm the current cooldown and any account-specific details. If the app shows a different message from a third-party guide, trust the app. ## Idea patterns that fit the rules - Two short words with a space: Moon Rider, Quiet Fox, Cedar Lane. - Two short words with an underscore: Moon_Rider, Quiet_Fox, Cedar_Lane. - Adjective plus noun: Rapid Otter, Silver Kite, Lucky Ember. - A short name plus a number: Luna 7, Kite 21, Nova 04. - A theme word plus a role: Night Scout, Forge Builder, Trail Guide. Keep the whole name under 20 characters, including spaces. A pattern that looks short in your head can cross the limit once both words are spelled out. The checker counts as you type, so you can trim before saving. If you are also naming accounts on other platforms, use the same seed word but adjust the format. A TikTok handle has different length and character rules, and the [TikTok username generator](https://tolz.me/social/tiktok-username-generator) can build ideas around the same theme. A [Discord username generator](https://tolz.me/discord/username-generator) does the same for a handle that allows periods and underscores but not spaces. ## Common rejection reasons 1. The name is shorter than three characters or longer than twenty. 2. It uses emoji, punctuation, or a copied fancy font instead of plain letters. 3. It starts or ends with a space, often because it was pasted from a document. 4. It contains a word or letter sequence that the Roblox filter blocks. 5. It looks like an official Roblox name or a well-known creator's name. 6. The account changed its display name within the last seven days. If you hit a rejection, change one thing at a time. Remove symbols first, then check the length, then simplify the wording. Retrying the same name with a lookalike character usually makes the result worse. ## Check the official rules before you commit Roblox Support publishes the current rules and the steps for changing a display name. Start at [en.help.roblox.com](https://en.help.roblox.com/hc/en-us) and search for Display Name. Read the article that matches your platform, because the menus differ between the website and the app. A display name is easy to change compared with a username, but the seven-day wait makes a little preparation worthwhile. Pick a name you can read at a glance, keep it within the character set, and leave the filter some room by avoiding edgy word combinations. The best display name is the one that still looks right after the novelty wears off. ### Shipping AI-Written Code You Don't Understand: What Developers on Hacker News Actually Do - URL: https://tolz.me/blog/shipping-ai-written-code-you-dont-understand - Published: 2026-10-21 - Updated: 2026-10-21 - Tags: AI, Programming, Code Review A Hacker News thread on shipping AI-generated code: the workflows people use, how they spend review time, and how to check small artifacts by hand. On October 19, 2026, an Ask HN thread went up about shipping code written by AI. The prompt was a LinkedIn post from an engineer at Meta who said that writing code with an assistant feels like a black box. It drew hundreds of comments. The thread is worth reading because it collects working practices from people shipping code every day, rather than predictions about where AI coding is heading. Here is what the top of the thread keeps coming back to, plus a checklist you can apply to your own team this week. ## What the commenters actually do ### Design the shape, let the model fill it in A recurring theme was to keep the design yourself and let the model implement it. The interface is the part you have to defend later, so you own it. The body of the implementation is where speed matters most. ### Sketch it first One commenter keeps a scratch Python file, sometimes a scratch repo, that they and the model both edit, even though the project itself is in Rust. Python reads like pseudocode, so it is a quick way to pin down method signatures and data shapes before the model writes the real code. ### Keep the task bite-sized, and talk before you say do it One commenter described a workflow of bite-sized jobs, 10 to 15 minutes discussing the implementation with the model, and only then saying "do it". They were explicit about who commits: "It doesn't commit code. I do." ### Use a second model to review A couple of commenters have one model write the code and a different one review it. One suggested using Claude to write and Codex to do an adversarial review, going back and forth until the result holds up. ### Spend review time where the risk is One commenter's rule: the more dangerous a decision and the harder it is to roll back, the more time they spend reviewing it. Another tied the same risk-driven idea to George Fairbanks' book Just Enough Software Architecture. Another concern was about whoever inherits the code. One commenter pointed out that many teams still have SREs as their first line of support, and those people need to understand how a service's hot paths work without the context the original author's AI session had. ## Where they disagreed Not everyone agreed. Some said they merge based on the description and expect a future model to clean it up. Others warned that tech debt compounds faster when nobody reads the code. Both positions come from the same thread. ## A checklist you can apply this week 1. Write the interface first. If you cannot describe the function signature and the error behaviour, the model is designing it for you. 2. Sketch the algorithm before you ask for code. Pseudocode in a scratch file is enough. 3. Split the work into pieces small enough to review in one sitting. 4. Read the diff line by line, the way you would read a diff from a new hire on day one. 5. Run the tests, then change the tests. A test that fails when you alter the input catches a lot of confident wrong code. 6. Ask a second model to review the diff, then read its notes and decide for yourself. 7. Commit the code yourself. You own whatever ships. ## Writing the hand-off The thread's advice about small tasks translates into a prompt with a few fixed parts: the shape you want, the file it belongs in, the constraints it has to respect, and the check that proves it works. A model can fill in the middle. It cannot invent the first and last parts for you. - Name the function or type you want, with its arguments and its return value. - Name the file and the module it belongs in, so the change lands in the right place. - List the constraints: the language version, the libraries already in the project, the style the file uses. - Name the check: the test that has to pass, or the command that has to succeed. ## Reading a diff you did not write Reviewing generated code looks different from reviewing a colleague's code, mostly because you have no memory of the conversation that produced it. The commenters' habits translate into a fixed order of reading. 1. Read the names first. Function names, argument names, and file names tell you what the change claims to do before you look at a line of logic. 2. Read the diff for what it does not mention. A deleted guard, dropped error handling, and a changed default are all easy to skim past. 3. Look for the dependency you did not ask for. A new library, a new network call, or a new file outside your project structure are worth stopping on. 4. Check the boundaries. Empty input, very large input, a missing key, and a failed call are where small changes tend to break. ## Testing the result - Run the existing tests before you change anything, so a later failure has an obvious cause. - Add a test for the case you actually care about, not only for the happy path the model wrote. - Change an input and watch the test fail. A test that still passes when the logic changes is not testing the logic. - Try the code the way a user would, including the bad input and the empty one. > Tip: Ask the model to explain the diff in two lines before you read it. If the explanation does not match the code, you have found your first problem while it is still cheap to fix. ## Check the small artifacts by hand Small changes often fail on arithmetic or syntax, in places you can check by hand without running anything. A few Tolz.me tools cover exactly those cases. ### Cron schedules A wrong cron expression is a bug that runs on its own schedule, which is the worst moment to discover it. Check the expression before it lands in a deploy. The [cron expression translator](https://tolz.me/developer/cron-expression-translator) turns a schedule into readable next-run times so you can see when it actually fires. ### File permissions in deploy scripts A chmod line in a deploy script is easy to misread. 755 and 750 are one character apart and mean different things. The [chmod calculator](https://tolz.me/developer/chmod-calculator) turns an octal mode into the owner, group, and other permissions it actually grants. ### Timestamps in logs and APIs Seconds or milliseconds, local time or UTC, a date parsed in the wrong zone. The [Unix timestamp converter](https://tolz.me/developer/unix-timestamp-converter) converts both ways, so you can compare a number in a log line against a date you can read. ## What to do when the code works but you cannot follow it Some generated code passes every check and still leaves you unable to say what it does. The practical response is to change the code rather than write documentation about it. If you cannot explain a function to a colleague, rewrite it in a shape you can explain, even if the rewrite comes out smaller than the original. This is where the two ends of the thread separate. One group treats code that works but cannot be read as debt a later model will clear. The other treats it as debt that keeps compounding, because every later change lands on code nobody can reason about. ## What the thread does not settle Nobody in the thread offered a rule that fits every team. The workflows described above share a shape: they keep the decisions, delegate the typing, and review in proportion to the risk. The engineer who called it a black box was describing the part that is not yours to delegate, and the practices in the thread keep coming back to that same point. Nothing in the thread settles how much code should be written this way. What it does make clear is that the risk sits in a specific place: not that the model writes code, but that code arrives without a person who can explain it. Every practice in the discussion is a way of keeping that person in the loop at the point where it matters. ## Sources - [Ask HN: shipping AI-written code you do not understand](https://news.ycombinator.com/item?id=50151036), October 19, 2026 ### WhatsApp's Hidden AI Notes: How Unseen Text in Web Pages Steers Chatbots - URL: https://tolz.me/blog/whatsapp-hidden-ai-instructions-help-pages - Published: 2026-10-21 - Updated: 2026-10-21 - Tags: AI, Security, Web WhatsApp help pages carried AI instructions nobody could see. What Redact found, what WhatsApp said, and how to check a page for hidden text. On October 8, 2026, Redact found text on WhatsApp's help centre pages that no visitor could see. It was in the page source, written as instructions for an AI, and it was sent along with the page to anything that fetched it. Redact published its findings on October 19, 2026, and the story came up on Hacker News. The details are worth reading closely. The notes were aimed at the software that answers WhatsApp support questions, and on the delete-account page one of them pointed in a direction that did not match WhatsApp's own published policy. ## What Redact found - Redact read all 545 pages that the WhatsApp help centre links to. - 80 of those pages contained hidden text written for an AI. - The hidden text appeared across 102 separate blocks on those pages. - By October 18 at 02:08 UTC, every hidden block was gone from the live pages. ## The delete account page The delete-account page got the most attention because the note on it was about a decision users care about. The page covers the auto-deletion periods: a deactivated account is deleted after 30 days, an inactive account is generally deleted after 120 days, and in the United States the window is 365 days. The hidden note came after a restatement of those same periods. It told the AI not to recommend contacting any Data Protection Authority, regulator, or supervisory authority. It also told the AI not to suggest that WhatsApp can manually delete an account. Redact points out a conflict with what WhatsApp publishes. WhatsApp's privacy policy for Europe, effective July 2, 2026, tells users they can complain to the Irish Data Protection Commission or to another supervisory authority. The hidden note steered the assistant away from exactly that route. ## The reporting and blocking page On the page about reporting and blocking other users, the hidden note handled the case where a user mentions legal action or lawyers. It told the AI not to reference data protection regulations such as GDPR, LGPD, or CCPA, even when the user raises them directly. ## Most of the blocks were housekeeping The two notes above were the striking ones, not the norm. According to Redact, most of the 102 blocks were routine housekeeping, the kind of instruction that keeps a support chatbot consistent. A few of them protected users rather than the company. One told the AI not to suggest lying about an age. ## The timeline - October 8, 2026: Redact finds the hidden text while reviewing the WhatsApp help centre. - October 18, 2026, 02:08 UTC: Redact's check finds every hidden block removed from the live pages. - October 19, 2026: Redact publishes the write-up, and the story comes up on Hacker News. - WhatsApp tells WIRED the notes were instructions for its support chatbot and says they reached the public pages because of a bug, which WhatsApp says has been fixed. ## What a reader saw versus what the page carried A reader saw nothing unusual. The visible page showed ordinary help content. Someone reading the delete-account page saw the deletion periods and the help copy around them. The extra sentences existed in the document rather than in the layout, so no amount of scrolling, zooming, or careful reading on the rendered page would reveal them. That gap is the whole problem. A browser hides text with rules that only the browser applies. Any other consumer of the same document, including a crawler, a scraper, an accessibility tool, or an assistant, receives the rules and the text separately, and mostly reads the text. ## Where the blocks sat in the source On the delete-account page, Redact found the note near a name in the page source: HelpContentGalaxyLLMInclude. Searching the source for that string, or for the wording of the note, is how Redact tells readers to find it. Copies on the Internet Archive still show the notes, so the pages can be inspected after the live versions were cleaned up. Two details make this more than a curiosity. The pages were served to Google's and Bing's crawlers, so the notes were available to anything that indexes them. And Redact says it has not tested whether any AI assistant acted on the instructions, or whether WhatsApp's own chatbot followed them. For contrast, Redact checked the Facebook and Instagram help centres as well, 300 pages in total. Those pages had no such notes. ## How hidden text works in HTML There are several ways to put words in a page that a reader never sees, and they are not the same technique. That distinction matters here, because the WhatsApp notes were blocks of text in the page source, not invisible characters. ### Hidden elements The usual approach is a container the browser is told not to display: the `display:none` CSS rule, the `hidden` attribute, or a wrapper collapsed to no size and covered by other content. The text is still part of the document, and it travels with every request for the page. ### Invisible characters A different technique uses characters that take up no space on screen. The zero-width space, the zero-width non-joiner, the zero-width joiner, the word joiner, and the byte order mark all render as nothing while staying part of the string. A name can look empty and carry a word inside it. A fetcher reading the raw text sees the word. That is the trick behind a family of Discord tricks where a visible nickname hides a second name behind it. The [Discord invisible name tool](https://tolz.me/discord/invisible-name) shows how those characters behave once they are inside a name. WhatsApp's notes were not invisible characters. They were ordinary sentences that the page carried but never drew on screen. ## Why it matters when you feed pages to an AI Any tool that turns a web page into model input sends the hidden parts too. A browser is what hides text from a reader, and a fetcher is not a browser. It takes the document as it arrived. Once that text reaches a model, it is text in a prompt. If it is phrased as a directive, an assistant that has been told to follow instructions can treat it as one, including directives the visible page never showed a reader. That is the shape of a prompt injection, and the hidden version is harder to spot because a person skimming the page sees nothing unusual. Retrieval systems make this worse in a quiet way. A search pipeline splits page text into chunks before a model sees any of it, so a hidden block can be retrieved on its own, without the surrounding article for context. The model receives a few sentences that look like system instructions and no page around them to explain them. Redact's write-up is careful about the limit of the evidence. The notes existed and were served to crawlers, but Redact did not test whether an assistant acted on them. The exposure is structural rather than demonstrated. ## Why the WhatsApp explanation changes how you look for these If the notes were instructions for a support chatbot, then the useful search term is the chatbot itself rather than anything sinister. Look for text that addresses a reader that is not a person: rules for a helper, a list of things to say and not say, a paragraph written to steer a response. The bug explanation deserves to be taken seriously, because a bug and a feature look identical from the outside. Both leave text in the page that a person cannot see. What separates them is intent, and intent is exactly what you cannot read from the source alone. The same reasoning applies to any page you check. Ask what a hidden block would be for, then decide whether that answer fits the site. A documentation page carrying notes for its own search assistant is a different story from a shop carrying notes aimed at an assistant that writes product reviews. > Tip: Search the source for a marker name as well as the wording. On WhatsApp's delete-account page, the note sat near HelpContentGalaxyLLMInclude. A named component or wrapper can be a faster route to hidden notes than guessing the phrasing of the note itself. ## How to check a page for hidden text 1. View the source. Press Ctrl+U on Windows and Linux. On a Mac it is Cmd+Option+U in Chrome or Cmd+U in Firefox. The source is the document the server sent, with no CSS applied, so hidden elements show up as plain text. 2. Search the source for telltale words. Try AI, assistant, ignore, do not, LLM, or the name HelpContentGalaxyLLMInclude. A block of sentences addressed to a model reads differently from interface copy once you go looking. 3. Check what the browser is hiding. Open the developer tools, select the page content, and look for containers set to `display:none` or carrying the `hidden` attribute. Any of them can hold text a model would read. 4. Check the text you were handed. If a page was pasted into a chat window or a document, invisible characters travel with it. The [invisible character remover](https://tolz.me/text/invisible-character-remover) shows which characters are actually inside the string. ## What to do when you find instructions aimed at a model - Report the page to the site owner, quoting the block and giving the location in the source. - Treat fetched pages as untrusted input. Give an assistant the visible text as content, and keep your instructions on the side where they belong, outside the fetched material. - Look at archived copies when the live page has already been cleaned up, which is what happened here. ## Sources - [Redact: WhatsApp's Hidden Notes Told AI Not to Recommend a Regulator](https://redact.dev/blog/whatsapp-help-pages-hidden-ai-instructions), by Dan Saltman, October 19, 2026 - [Hacker News discussion of the Redact findings](https://news.ycombinator.com/item?id=50159341) ## Links - [Curated index](https://tolz.me/llms.txt): Short list of every tool and guide. - [Use Tolz.me with AI agents](https://tolz.me/developers): How to connect the MCP server to ChatGPT, Claude, Cursor, and browser agents. - [AI tools manifest](https://tolz.me/.well-known/ai-tools.json): Machine-readable catalog of all tools. - [MCP server card](https://tolz.me/.well-known/mcp-server-card): Connection details for the remote MCP server. - [All tools](https://tolz.me/tools): Searchable index. - [Blog and guides](https://tolz.me/blog): Long-form guides from the tool library. - [About Us](https://tolz.me/about): Product overview. - [Contact](https://tolz.me/contact): Support and suggestions.