# Panella > Governed, self-hosted memory for AI agents. A standard MCP server where every write is proposed, > approved by a named person, and made durable only against a chain-verified approval receipt. > Default-deny, fully auditable, Apache-2.0, runs on your own box. ## For agents - [Agent install guide](https://github.com/panellatech/panella/blob/main/llms-install.md): install and onboard a box end-to-end (prerequisites, `uvx panella up`, client wiring, verification, operator hand-back) ## Docs - [README](https://github.com/panellatech/panella/blob/main/README.md): product overview and quickstart - [Quickstart](https://github.com/panellatech/panella/blob/main/docs/QUICKSTART.md): `panella up` to first approved memory — one person, one box, ~15 minutes - [Team recipe](https://github.com/panellatech/panella/blob/main/docs/recipes/claude-code-team-memory.md): one governed box for a small Claude Code team — install, teammate on-ramp, daily approval rhythm, offboarding - [Self-hosting](https://github.com/panellatech/panella/blob/main/docs/SELF_HOST.md): configuration and Docker topology - [Governance](https://github.com/panellatech/panella/blob/main/docs/GOVERNANCE.md): the approval and audit model - [How Panella compares](https://github.com/panellatech/panella/blob/main/docs/COMPARISON.md): the two branches of agent memory and where Panella stands — category-factual, no vendor named - [Security](https://github.com/panellatech/panella/blob/main/SECURITY.md): security posture and reporting - [Console](https://github.com/panellatech/panella/blob/main/docs/CONSOLE.md): flag-gated operator console - [Upgrade](https://github.com/panellatech/panella/blob/main/docs/UPGRADE.md): backup, upgrade, rollback - [Tenancy](https://github.com/panellatech/panella/blob/main/docs/TENANCY.md): multi-tenant boundaries ## Optional - [Release runbook](https://github.com/panellatech/panella/blob/main/docs/RELEASE.md): how releases are built, signed, and verified - [Contributing](https://github.com/panellatech/panella/blob/main/CONTRIBUTING.md): development setup and expectations