> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.recursiv.io/guides/govern/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.recursiv.io/_mcp/server.

# Govern

> Quickstart, SDK, MCP and REST API reference. Build, run and govern autonomous agents on Recursiv.

## Docs

- [Govern Overview](https://docs.recursiv.io/guides/govern/overview.md): Run agents on Recursiv so you can govern and prove what they did
- [Identity & Access](https://docs.recursiv.io/guides/govern/identity-access.md): Every agent and key has a scoped identity, with least privilege by default
- [Guardrails & Approvals](https://docs.recursiv.io/guides/govern/guardrails-approvals.md): Per-tool policy and a human approval queue for sensitive actions
- [Audit & Observability](https://docs.recursiv.io/guides/govern/audit-observability.md): Read the record of what your agents actually did
- [Evidence](https://docs.recursiv.io/guides/govern/evidence.md): Measure what your agents actually did — and prove it.
- [Self-Evaluation](https://docs.recursiv.io/guides/govern/self-evaluation.md): The Recursiv Protocol, agents grade their own output and refine before delivering
- [Agent Trust Architecture](https://docs.recursiv.io/guides/govern/agent-trust-architecture.md)