What it does
Brain is one of the open-source agents built on Agent-Native, a TypeScript framework for agents that pair autonomous work with a purpose-built UI. The publisher describes it as citing company knowledge from Slack, meetings, transcripts and decisions. In the framework, each capability is defined once as an action that the agent can call as a tool and the UI can call from code, using the same validation and permissions. The agent also receives relevant UI state, such as the current page or selected record.
The framework suits teams that want people and agents to work on shared data in one interface. It includes agent chat, authentication and permissions, skills and memory, automations on schedules or events, and agent teams. Users can review, edit, approve and share agent work in the UI.
Running it requires your own LLM and SQL database. Production uses PostgreSQL, while local development can use PGlite, on any Nitro-compatible host. Visual editing of a running app is driven from Claude Code using the /visual-edit command. Contributions are described in CONTRIBUTING.md, and DEVELOPMENT.md covers local setup.
Limitations and requirements
Requires your own LLM and SQL database. Production uses PostgreSQL, and local development can use PGlite, on a Nitro-compatible host. The visual-edit workflow is run from Claude Code.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
Agent-Native Brain
Cited company knowledge from Slack, meetings, transcripts, and decisions
v1.0.0As shown on registry.modelcontextprotocol.io when we last read it