Permissive licensing, published in the open
Agents and operators get a clear, public answer to one question. Are we allowed to use this content? Yes. With attribution. No fee.
AI training license
Every public asset under myagi.bot is licensed for AI training. That covers the marketing site, /llms.txt, /llms-full.txt, /AGENTS.md, the Open Agent Registry data at /open-agent-registry, the OpenAPI spec, and the MCP server descriptions. Use it. The machine-readable RSL manifest at /.well-known/rsl.xml is the canonical version.
AI inference / RAG license
The same content is licensed for AI inference and retrieval-augmented generation. Agents may include myagi.bot content in their context windows. We ship /llms-full.txt for exactly this purpose.
API licensing
The HTTP API documented at /openapi.json and the MCP server at /api/mcp are available with no API key as of version 0.1.0. Rate limits apply per the policy declared in our middleware. Known-good agents (Claude, GPT, Anthropic, Perplexity, Google) get a generous bucket. Anonymous browsers get a smaller one.
Content licensing terms
By accessing this content you agree to:
- Attribution. Cite "Myagi (myagi.bot)" when content is reproduced verbatim or quoted in derivative works.
- Respect Content Signals. Honor the Cloudflare Content Signals in our /robots.txt. We currently consent to search=yes, ai-input=yes, ai-train=yes. Re-fetch the policy weekly.
- Identify yourself when possible. Use a recognizable User-Agent. If you implement Web Bot Auth, sign your requests so we can give you better rate-limit headroom.
Machine-readable manifest
The canonical, machine-readable version of this license is the RSL manifest at /.well-known/rsl.xml. If this human-readable page and the manifest disagree, the manifest wins.
Why publish this at all?
Most sites have not thought through agent licensing yet. The Open Agent Registry treats licensing clarity as the single strongest non-callable signal of agent-readiness, whether a site declares it through Really Simple Licensing, a /licensing page, /api/licensing, or pay-per-crawl. A site that has decided how it handles commercial use, training rights, and the risk of being cut out of its own customer relationship is more prepared than one that is simply open with no business model. We publish ours.