# Make your website readable by AI agents. Source: https://agentreadable.site/ [NEW The Pons Readability Index, monitor alerts and a CLI](https://agentreadable.site/pons/) Paste a URL. Readable reads up to 30 public pages, strips menus, scripts and banners, and writes an llms.txt plus clean markdown copies for you to upload to your own site. [Check a score](https://agentreadable.site/check/) - No login - No wallet - Public pages only - Up to 30 pages - robots.txt respected - Free An llms.txt as Readable writes it, for an illustrative token site. The files work wherever you can upload them - Next.js - Vite - Astro - SvelteKit - WordPress - GitHub Pages - Netlify - Cloudflare Pages - Any static host The difference ## From unreadable to agent-ready in _one_ upload. Agents skim. A site built for browsers hides its text behind menus, scripts and banners. Readable writes the version agents read in one request, and shows the score it will get before you upload anything. - Every check of the score is explained, with the pages that fail it. - The score after upload is projected honestly, never padded. - Pages that need JavaScript are flagged, not guessed. - llms.txt - llms-full.txt - pages/\*.md Example scan. The sites and figures are illustrative. How it works ## Four steps, no account. Readable only reads what a visitor could read. You stay in charge of your site: it never asks for a login, a wallet or access to your hosting. 1. ### Paste your URL Readable reads your robots.txt first, then your homepage. Login and admin pages are never touched. robots.txt → homepage 2. ### We read and clean Up to 30 pages, closest first. Menus, scripts and banners go; the text and its structure stay. 30 pages · 70–95% smaller 3. ### Download and upload One zip with llms.txt, llms-full.txt and a markdown file per page. Put it in your site's web root. readable.zip → /public 4. ### Check your score The checker reads your site the way an agent would and scores it from 0 to 100, with a badge and daily monitoring if you want them. /check · /badge · /monitor What a scan keeps ## Noise out, text in. Every page is edited down to what a reader needs, then written as markdown that agents can read in one request. Struck out - navigation menus - scripts - cookie banners - pop-ups - hidden elements - tracking pixels - footers - share buttons Kept - headings - paragraphs - lists - tables - code - links - image text 70–95% smaller than the HTML, page by page 30 pages per scan, closest to your homepage first 0 logins, wallets or cookies | File | What it is | | --- | --- | | `llms.txt` | Site title, one-line summary, and your pages grouped into About, Tokenomics, Docs and Optional. | | `llms-full.txt` | Every page's markdown in one file, for agents that want everything at once. | | `pages/*.md` | One clean markdown copy per page, linked from llms.txt. | The score ## One score, from 0 to 100. The checker fetches your `/llms.txt` and the pages it lists, the way an agent would. Five checks, the same rules everywhere. [Check any site →](https://agentreadable.site/check/) 1. An llms.txt with a title and sections At /llms.txt, with an H1 title and at least one section. 20 pts 2. A summary under the title One line in a blockquote: what the site is. 15 pts 3. Markdown copies of the listed pages Proportional: each listed page counts. 25 pts 4. Text without JavaScript The listed pages show their content as served. 20 pts 5. Links that resolve Every link in llms.txt answers 2xx. 20 pts **Not readable** 0–49 **Needs work** 50–79 **Agent-readable** 80–100 Read this first ## What Readable can't promise. llms.txt is young. Here is what it does today, and what this tool is not, before you rely on either. Not guaranteed today - AI products are not required to read llms.txt. Some agents and developer tools fetch it when they find it; many don't. We can't promise any specific AI will use it. - llms.txt is a community proposal ([llmstxt.org](https://llmstxt.org/)), not an official standard. The format can still change. - It is not an SEO tool. Don't expect search ranking changes. - Pages that only show content after JavaScript runs can't be converted. Readable flags them instead of guessing. What this is not - Not a security audit, and not a safety check for any token. - Not investment advice. - Not hosting: you download the files and upload them yourself. - Not a browser: no headless rendering, no JavaScript execution. - No login, no wallet connection, no access to your hosting. More tools ## Keep it readable. The same checks, wherever you need them: on a page, in your README, in your inbox, or in CI. [Checker](https://agentreadable.site/check/) Score any site Five checks in a few seconds, with what to fix for each one. [Badge](https://agentreadable.site/badge/) Show your score A live SVG badge for your README or site footer. [Monitor](https://agentreadable.site/monitor/) Know when it breaks A daily re-check with Telegram or webhook alerts. No account. [Index](https://agentreadable.site/pons/) Pons Readability Index How readable are the websites of tokens launched on Pons? Ranked weekly. [Command line](https://agentreadable.site/docs/#cli) npx agentreadable Scan and check from your terminal, or fail a CI build below a score. [Docs](https://agentreadable.site/docs/) How it all works The score, the crawler, the upload, privacy and limits. ## Make your site _readable_. One URL, a few seconds, and a zip of files you can upload today. Free. No account. Public pages only.