Documentation
Documentation
Welcome to the Agentgram documentation.
Agentgram is a unified interface for interacting with multiple AI agents. The API is a Go multiplexer that connects to remote agents over their native protocol (Custom REST/SSE, A2A or Google ADK), handles authentication, permissions and session storage, and converts everything to a uniform stream of AG-UI events. The web client renders that stream.
Centralize everything, behind RBAC. Register every agent (ADK, A2A or custom) and every MCP
server your company runs, and reach them all through a single API endpoint and a single MCP endpoint.
Role-based access control decides — per agent and per MCP server — exactly who can access what, so one
integration gives each user only the tools they’re entitled to.
Get started
One compose file. A demo agent. Chat in under two minutes.
Self-host API + web + Redis + PostgreSQL on a server.
Deploy with Helm using the bjw-s app-template chart.
Learn the concepts
The config file, environment variables and secrets.
Register agents and how each protocol maps to AG-UI.
Share conversations and create shared multi-agent groups across users.
Expose your agents as tools in any MCP-compatible IDE or CLI.
Vs Open WebUI, MCPJungle, ContextForge, agentgateway.
The full OpenAPI reference for the HTTP API.