Does it exist, and where?
We read the well-known addresses where sites publish files for agents. Our crawler identifies itself, honours robots.txt, and publishes its addresses.
Neutral public index
IntentRoute is a neutral public index of the files that websites publish for AI agents: UCP profiles, A2A agent cards, MCP server descriptions, llms.txt, OpenAPI, ARD and AI Catalog, APIs.json. We check whether they exist, whether they are valid against the protocol they claim, and how they change over time. Every fact carries its source, a verbatim quote, and the date we read it.

What we do
We answer the same three questions for every provider and every profile, with the evidence attached.
We read the well-known addresses where sites publish files for agents. Our crawler identifies itself, honours robots.txt, and publishes its addresses.
Each file is checked against the official schema of the protocol version it claims, and the exact deviations are listed.
An append-only change ledger records every change we confirm, with dates and attribution. Nothing is rewritten.
Development stage
Live
Search APIs for agents, described claim by claim — price, limits, authentication, terms, privacy — each with its source, a quote, and the observation date, re-checked every week.
Change ledger at /changes, with JSON, JSONL, and Atom feeds.In progress
Our crawler reads the profiles websites publish for agents and grades each against its protocol. The first open dataset and report follow once the history covers several weeks.
Crawler policy, addresses, and opt-out at /crawler.Built for
Read sourced facts and their dates through machine-readable paths, including what is still unknown.
Read the resource map →Use the OpenAPI contract. No key is needed to read. Web-search provider records and the change ledger are CC BY 4.0; per-site data and profiles from 15 October 2026 are CC BY-NC 4.0, commercial reuse licensed.
Developer notes →See how your profile and your facts are read, and ask for a correction if we got something wrong.
How we describe providers →Principles
Open: every page, history and report is free to read for everyone, without a key (CC BY 4.0). Web-search provider records and the change ledger are CC BY 4.0. Per-site data and profiles published from 15 October 2026 are free for non-commercial use (CC BY-NC 4.0); commercial reuse is licensed.
Neutral: we do not sell placement, rank providers, or route or resell traffic.
Correctable: every fact links to its source; corrections are added as new events, never by rewriting history.
View the roadmap