# Access

Every page and file on Agent Ready Crawlers is public. There is no account, API key, token or login. Nothing is gated. This file states that for each step an agent might expect to take.

## Discover

Start at / or /llms.txt. All URLs are in /sitemap.xml. No discovery step needs a credential.

## Pick a method

Fetch pages as HTML or as markdown (add index.md to the path, or send Accept: text/markdown). Datasets are plain CSV and JSON files under /data/. There is one method: an ordinary request with no credential. Pages and files take GET; the MCP server at /mcp takes JSON-RPC over POST; the REST API under /api/v1/ and the NLWeb endpoint at /ask take GET, or POST with a JSON body.

## Register

Not required. There is nothing to register for.

## Claim

Not applicable. No credential exists, so there is nothing to claim.

## Exchange

Not applicable. No token is issued, so nothing is exchanged. The site publishes no OAuth metadata (no /.well-known/oauth-authorization-server and no /.well-known/oauth-protected-resource) because it has no authorization server: no endpoint, the MCP server and the API included, takes a token or a key. The machine-readable form of this statement is `security: []` with no security schemes in /openapi.json.

## Use

Send a plain GET with an honest user-agent string. The site's robots.txt allows the AI crawlers it writes about. Fetch the datasets from /data/, or query them through /mcp, /api/v1/ and /ask as /developers describes. Cache by the dates on each page.

## Errors

A missing path returns 404 directly; under /api the 404 is JSON with a code, a message and a hint. A server error returns 5xx; retry later. The API and the MCP server allow 120 requests a minute per address, announced in RateLimit headers, and a 429 carries Retry-After. The host may add its own 429 above that. Wait that long and try again.

## Revocation

Not applicable. There is no credential to revoke and no access to withdraw.