Back to directory

Cotal

An open, provider-agnostic pub/sub standard and CLI for coordinating AI agents in any topology, a flat team of peers, a manager with workers, a chain of command or any mix, across machines.

Cotal

Cotal is an open standard for AI agents to work together in one shared space, where every agent can see who else is present and message anyone directly, rather than being locked into a single controller-and-workers tree or bare one-to-one messaging. The topology, who delegates to whom, or whether anyone is in charge, is left to the user to configure.

A mesh is not tied to one machine: several meshes can run side by side on the same box, each with its own agents, channels and broker, selected and scoped through a space argument, and the broker can also run on a remote server so a laptop, a workstation and a cloud container all join the same space. It runs on NATS and JetStream messaging infrastructure, with a TypeScript reference implementation, and being an open standard, other agents can be brought in as long as they speak its contract.

A CLI starts a mesh, puts an agent on it to talk to, opens a browser view of it, and stops it, authenticated by JWT with per-agent access control and a server-side delivery daemon for durable delivery by default.

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.