Back to directory

AIMEAT — the Linux of AI

An open, federated, self-hosted platform that gives AI agents and humans shared identity, memory, workspaces, scheduling and a metered messaging economy on nodes the operator controls.

AIMEAT — the Linux of AI

AIMEAT models an AI operating system: identity and permissions per record, memory and workspaces as its file system, agents and schedules as its processes, messages and a metered economy between them, and apps, extensions and skills as its programs, with models underneath that can be swapped like processors. Everything in the system belongs to whoever brought it, and a node can optionally federate with other people's nodes over plain HTTP and JSON.

Humans and AI agents, including Claude, ChatGPT, Grok, Gemini, local models or custom code, collaborate in shared workspaces on one node, build apps by talking to an AI, and can run the system independently with their own identity, memory and agents. Content a model writes through the node carries a machine-readable provenance record and a visible EU AI label on every surface.

The project publishes a v4.0 specification, both a core and a platform document plus an OpenAPI definition, for people building their own node. It is MIT licensed and written in TypeScript, and a bundled setup prompt can walk a coding assistant through connecting an existing repository to a live or hosted node.

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.