Back to directory

Agents Universe

An open-source enterprise AI agent platform where agents onboard by reading project documentation, plan tasks, and execute them under their own identity and permissions, leaving an audit trail.

Visit website (opens in a new tab)agent.agents-universe.com
Agents Universe

Agents Universe is an open-source (Apache-2.0) enterprise AI agent platform built to make agents do real work rather than function as another chat interface. It ships with ten ready-to-use agents, including a Product Owner, a Tech Lead, a Quality Assurance agent, an office assistant that turns content into deliverables, and an agent that runs authorized white-box penetration tests within scope.

Agents onboard by reading the full project documentation instead of using vector retrieval, plan before taking on a task, execute under their own identity and permissions, and leave an auditable trail. Experience gained from documents or completed work is distilled back into skills, workflows and knowledge assets that the whole project can reuse. Built-in agents can be brought into a conversation with an @-mention and coordinate with each other, and each user configures their own LLM provider and API key.

The backend is Python and the frontend is Vue 3, and it self-hosts with a single Docker Compose command, with PostgreSQL, MySQL, SQLite and SQL Server database support.

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.