Back to directory

Agentgram

A single gateway that unifies REST, A2A, and Google ADK agents behind one API and one MCP endpoint, converting every response into a uniform stream of AG-UI events with role-based access control.

Agentgram

Agentgram acts as a front door for an organization's AI agents and MCP servers: it multiplexes authentication, permissions, and session storage, talks to each backend agent in its native protocol, and converts every response into a uniform stream of AG-UI events so a single web client can render any agent's output the same way, regardless of whether it speaks REST, A2A, or Google ADK.

Role-based access control decides which agents or MCP servers each user or group can reach, so one integration can safely expose many internal agents. Conversation history persists per agent in Redis while the agents themselves remain stateless, and the project adds shared, moderated group chats where multiple agents and people collaborate, plus revocable, time-limited links for sharing a conversation.

More in Agent frameworks

See all 902
  • Recurse

    recurse.run

    A serverless harness that lets a coding agent design, test, and deploy its own specialist sub-agents as tools, MCP servers, or bots, defined by a schema-validated manifest.

  • ProofForge

    shulgin.is-a.dev

    An AI-agent pipeline that decomposes a mathematics problem, proves each piece, and formalizes the result in Lean 4 and Mathlib until the Lean kernel accepts it, with six of its proofs already merged into Google DeepMind's formal-conjectures repository.

  • local-coder

    github.com

    Builds a local coding-agent team on top of Ollama and OpenCode by detecting the machine's hardware, recommending models for seven specialist roles, and testing whether each assigned model can actually use tools, edit a repository and pass verification.