Developers

The catalog is a database with a website in front of it, and the database is readable.

REST API — the public endpoints: list and filter tools, read one, read the facet taxonomy. No key required.

MCP server — how our editors read and write the catalog from an AI client. Staff-only, but the design is the interesting part.

Plain text, for readers that are not browsers

Two files, in the shape llmstxt.org describes:

• /llms.txt — the index: every listed tool with its price in one line, plus links to the reference pages.
• /llms-full.txt — every fact the catalog holds about every listing: vendor, entry price, free tier, trial, facets, price provenance, the date we last checked, and the vendor's own pricing page.

Both are generated from the same list response the site renders, so they cannot drift from what a reader sees. If you are pulling this catalog into a model's context, start here rather than scraping the HTML.

Licensing and attribution

The prices and the taxonomy are facts. The editorial writing — descriptions, strengths, limitations — is ours. If you republish figures from here, please link the tool's page so a reader can check the provenance and the date, which are the parts that make the number worth anything.