# canireach

> A behavioral access map of the agent web. We measure whether agents can reach the top 1000 domains, what stops them, and which sites publish a detour. Numbers are recomputed from published capture bytes; every capture is downloadable with its SHA-256.

## Data

- [https://canireach.ai/api/summary.json](https://canireach.ai/api/summary.json): headline figures, per-caller reachability, affordance adoption, challenge vendors
- [https://canireach.ai/api/baseline.json](https://canireach.ai/api/baseline.json): the full derived aggregate for the current capture
- [https://canireach.ai/method](https://canireach.ai/method): measurement contract, rails, denominator discipline, comparability rules
- [https://canireach.ai/bot](https://canireach.ai/bot): what our probe is and how to block it

## Source

- [https://canireach.ai/](https://canireach.ai/): the human-readable map
- [https://github.com/BrunoOGclaw/canireach](https://github.com/BrunoOGclaw/canireach): instrument source, tests, and issue tracker
- [https://github.com/BrunoOGclaw/canireach/releases](https://github.com/BrunoOGclaw/canireach/releases): immutable captures with manifests and hashes