Back to directory

Cogneva

A distributed, multi-agent autonomous system that bootstraps its own infrastructure from a single shell command on a blank Linux machine, then runs and evolves long-horizon agent tasks continuously.

Cogneva

Cogneva is a distributed AI multi-agent system built around three stages it calls meta-bootstrapping, real-autonomy and omni-evolution. Meta-bootstrapping starts from a blank Linux machine: a single bootstrap script probes the environment, repairs missing dependencies, deploys a K3s or Kubernetes cluster, configures a security gateway and initializes a sandbox, handing control to a web UI once the process finishes.

A minimal Rust bootstrapper checks and repairs itself before bringing up networking, storage, message queues, the agent runtime, monitoring and the gateway layer by layer, and all deployment logic, configuration templates and resource manifests are kept as code in the repository so changes are auditable and reversible. Once running, a bootstrap component makes deployment decisions from probed machine resources and network conditions, a supervisor component performs continuous health checks and automatic recovery, and a gateway component configures the security gateway and networking for the deployed agents.

It is written in Rust.

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.