Open-source platform for orchestrating, governing, and safely automating AI agents across your infrastructure.
The security-first platform for orchestrating, governing, and safely automating AI agents. Default Deny. Explicit Allow. Always Auditable.
Overview · Quick Start · Features · Architecture · Install · Configuration · Screenshots · License
Akili is a platform for orchestrating, governing, and safely automating AI agents. It lets AI agents write code, operate servers and drive deployments, without handing them the keys.
- Orchestrate: hand work to a fleet of agents from one control plane — coding tasks, alert triage, deployments and chat sessions — and watch them run in real time. - Govern: every model call goes through the control plane's gateway, and every tool call is checked against a signed policy twice — once on the agent and once on the control plane — before anything runs. Risky actions wait for a human. - Automate safely: agents run on your servers and dial out to the control plane; approved change plans run step by step and roll back automatically if a check fails, and every decision lands in a hash-chained audit log.
- server: the control plane, the akili binary. A web UI, a REST API with OpenAPI docs at /docs, and an LLM gateway that holds the provider keys. State lives in PostgreSQL; Redis carries events, presence, leases and leader election. - agent: the akili-agent binary. It runs on your servers, dials out over an encrypted tunnel (yamux over WebSocket) and runs chat and task sessions within the signed policy it receives. It holds no model or forge credentials, and is installed from GitHub Releases. - proto: the wire contract both sides import: envelopes, t
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/goakili--akili.html)
Reliability + security proxy for the Model Context Protocol (MCP): self-healing connections, runtime tool-security (rug-pull/poisoning detection), and a compliance-mapped audit trail.
sainitish1609/mcp-guard🛡️ Ultra-fast local security firewall, secret sanitizer, and context token compressor for Claude Code, Cursor, and MCP AI agents.
CTRLRun/ctrlrunThe execution safety layer for AI agents.
graphersal/graphersalIn-memory graph database and Gremlin traversal engine in Rust: browser playground, CLI, MCP server for AI agents, durable store, Python bindings.
slxca/opencontextA Model Context Protocol server for persistent project-specific AI agent context.
NousC/opennousThe context graph for agentic GTM teams. Unify the data scattered across your GTM tools into one identity-resolved account any agent reads in a single call.
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.