llms.txt validator
Use this llms.txt validator to inspect a pasted or published file. Review Markdown structure, llms-full.txt and optional link checks before editing it.
An llms.txt validator checks a file against the proposed format at llmstxt.org: one H1 title, an optional summary and H2 sections containing Markdown links. Paste text or request a site check to see errors and warnings by line. Link checks can reveal missing pages, but a valid file does not guarantee assistant use or citations.
What this format can establish
Checked against the llmstxt.org proposal v2, updated 10 August 2026.
The proposal describes a file that tools can explicitly load for context. It does not promise that consumer assistants automatically request it, use it for training or cite your site. OpenAI and Anthropic document crawler controls separately. Publishing a valid llms.txt does not grant access blocked by robots.txt or authentication.
The proposal lists the VS Code PagePilot extension as an integration that loads external context, and an MCP server for fetching llms.txt files. It links files published for OpenAI, Anthropic and Gemini developer docs. These examples do not establish automatic use by every consumer assistant.
Read the format proposal · OpenAI crawler documentation · Anthropic crawler documentation
How this tool works
- Choose text or a site
- Pasted text is checked locally unless you request link checks. Enter a site, directory or exact llms.txt URL. Site mode requests llms.txt and llms-full.txt in that directory. Relative destinations are valid; supply their directory or file URL to resolve them for network checks.
- Fix the reported lines
- The format proposal requires a title. The blockquote summary, introduction and sections are optional. Inside a section, use Markdown link lists with optional notes. An Optional section identifies material that can be skipped when a shorter context is needed.
- Treat live checks as a snapshot
- Every unique link is listed. Each batch checks up to 50 links with four requests at once, a 20-second budget and robots rules. Continue to the next batch to check remaining links. HEAD has a bounded GET fallback. Restrictions, timeouts and private destinations remain explicit.
Before you start
Check what assistants can find
Crawler access is one part of AI visibility. Run a free scan to see how assistants describe your business and which sources they cite.