mateclaw vs qm
Self-hosted Java agent platform: ReAct plus plan-and-execute 'digital employees' with an LLM Wiki knowledge layer, skills/MCP/ACP tools, approval-gated actions and eight IM channels in one JAR. — versus — Multiplayer agent harness for startups: every employee gets a scoped workspace — memory, files, keychain, crons, sandbox — in Slack and web, with Pi/OpenCode/Codex/Claude Code swappable underneath.
qm gives every employee a scoped agent workspace in Slack with Claude Code/Codex/OpenCode swappable underneath; MateClaw ships its own harness, knowledge layer and consoles instead of wrapping coding CLIs — though ACP lets it mount them as employees.
| mateclaw | qm | |
|---|---|---|
| Stars | 1.0k | 14k |
| Forks | 318 | 1.7k |
| Language | Java | TypeScript |
| License | Apache-2.0 | MIT |
| Last activity | today | 2 days ago |
| Topics | agents, orchestration, memory | agents, orchestration |
| Curated connections | 5 | 5 |
mateclaw — the curator's take
Pick it when the buyer is an IT department rather than a hacker: multi-user workspaces, RBAC and approval gates on sensitive tool calls, audit trail, Spring Actuator health, per-channel error isolation, and provider failover across DashScope/OpenAI/Anthropic/Gemini/DeepSeek/Ollama when a vendor 401s mid-request. Underneath it's a real harness — StateGraph runtime, Team Runs with a durable runId, task DAGs, human approval — not a RAG call in a wrapper. The cost is weight and gravity: Java 21 + Spring Boot 3.5 + Vue, an enormous and fast-moving feature surface (workflows, triggers, Wiki transformations, Dream consolidation, skill mining), and a China-first channel set (DingTalk, Feishu, WeChat Work, QQ) whose value drops sharply outside that ecosystem. For a personal agent on a $5 VPS this is orders of magnitude too much machinery.
qm — the curator's take
The org-level answer to "give everyone at the company an agent": per-person and per-channel scopes keep memory, credentials and files isolated, and the harness underneath (Pi, OpenCode, Codex, Claude Code) is swappable, so the deployment isn't a bet on one vendor. Security postures (strict/auto/dangerous) plus a hard command-deny policy that survives even Dangerous mode are unusually well thought through. NOT for individuals — this is Postgres + sandbox + deployment-repo infrastructure; a solo dev should run a harness directly. Quirk: contributions are accepted as prose ADRs, not code PRs.