Back to directory

Sympozium

Sympozium is a coordination layer for multi-agent AI systems on Kubernetes, giving agents shared situational awareness through selective permeability, structured handoffs and shared memory while each agent runs in an isolated workload defined as a Kubernetes CRD.

Sympozium

Sympozium runs each agent in its own isolated workload on Kubernetes, using a fleet of microVMs on nodes with KVM support, with every coordination policy expressed as a Kubernetes custom resource. It focuses on giving agents shared situational awareness, rather than better prompts, through structured handoffs and shared memory with selective permeability between agents.

An experimental AgentHarness component runs an administrator-approved external adapter in an isolated AgentRun, or a session-capable runtime for a private, continuing agent-to-chat conversation, while Kubernetes policy, identity, skills, memory, MCP access, observability and audit stay under platform control; this is described as an adapter boundary, not permission to run arbitrary upstream images.

It installs with Homebrew, a shell installer, or a Helm chart that requires cert-manager for webhook TLS, and is written in Go under the MIT license. It is under active development, with an unstable API.

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.