Back to directory

CowAgent

CowAgent is an open-source, self-hosted "agent harness" that plans tasks, runs tools and skills, and improves itself over time through memory and a growing personal knowledge base, reachable across the web and major chat platforms.

CowAgent

It decomposes a complex task and loops over tools until the goal is reached, and can form a team of agents that each keep their own role, model, skills and knowledge inside one shared conversation. Memory is kept in a three-tier structure, context, daily and core, with automatic distillation and hybrid keyword-plus-vector retrieval, and it auto-curates a Markdown knowledge wiki with a visual, browsable knowledge graph.

A self-evolution process reviews past conversations to improve its own skills, follow up on unfinished tasks, and consolidate memory and knowledge as it is used day to day. It reaches users through Web, WeChat, Feishu, DingTalk, WeCom, QQ, Telegram and Slack, and switches among model providers including Claude, GPT, Gemini, DeepSeek, Qwen, GLM, Kimi, MiniMax and Doubao from its web console.

Python, MIT licensed.

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.