🦞 MoltProbe

Can AI agents read and use your site?

Free, and no registration. Results in about five seconds.

MoltProbe checks any site for llms.txt, skill files, crawler access, structured data and agent-legible content, then scores it 0 to 100.

What the probe checks

Agent files
llms.txt, skill.md, an MCP manifest, ai.txt. The files agents read first to learn what your site offers.
Crawler access
Whether robots.txt lets GPTBot, ClaudeBot and PerplexityBot in, or shuts them out by accident.
Structured data
JSON-LD, title, meta description and OpenGraph tags on the homepage.
Legible content
Real text in the initial HTML, plus a sitemap and a feed. Pages that are empty until JavaScript runs are invisible to most agents.

Why this matters

A growing share of your visitors are not people. AI search engines like ChatGPT, Claude and Perplexity read pages to answer questions, and agents are starting to shop, book and sign up on behalf of the humans they work for. If those systems cannot read your site, you simply do not exist in the answers they give.

Most sites were built for a browser and a person looking at it. That is a different reader from a crawler that does not run JavaScript, or an agent that needs machine readable prices, structured data and a clear description of what you offer. A site can look perfect to a person and be almost invisible to software.

MoltProbe checks the signals that decide whether you show up: the files agents read first, whether your robots rules let them in, whether the important content is in the HTML rather than hidden behind a script, and whether your pages describe themselves in a way machines understand. It takes about five seconds and it is free.

Why the lobster? Molting is how a crustacean sheds a shell it has outgrown. The web is mid-molt: sites built for human eyeballs now get visitors that are software.