Back to directory

MateClaw

A multi-user, self-hosted personal AI agent platform built on Spring Boot, with multi-agent orchestration, MCP, skills and memory, automatic provider failover, and multi-channel bots such as Telegram, Feishu and DingTalk.

MateClaw

MateClaw is a self-hosted AI agent platform built on Spring AI Alibaba, aimed at multi-user, IT-department-manageable deployment rather than a single-person tool: it supports multi-user workspaces, approval gates on sensitive actions, a full audit trail, Spring Boot Actuator health monitoring, and per-channel error isolation so one connected chat platform's outage does not take down the others. It runs as one JAR, with persisted data and configured model, channel, and tool services controlled by the operator.

When a configured model provider fails, whether from an expired key, a 401 response, a network issue, or an exhausted quota, MateClaw automatically retries the current request against the next healthy provider in a configured priority order across DashScope, OpenAI, Anthropic, Gemini, DeepSeek, Kimi, Ollama, LM Studio, and MLX, parking failing providers in a cooldown window. Its agent runtime offers a native StateGraph loop (ReAct, Plan-and-Execute, Goals, Team Runs) or DeepSeek Harness as a managed external loop, both sharing the same conversations, workspace boundaries, tool guarding, and lifecycle controls, and it connects to Telegram, Feishu, and DingTalk bots.

It is written in Java.

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.