Mar 8
Loading video player...
ClawOps is an AI-powered incident response system built around group chat collaboration. It solves the problem of slow and manual production issue handling. Instead of just sending alerts, ClawOps helps detect, analyze, and fix errors within a shared group thread where humans and AI agents work together. Using Convos for conversational orchestration and XMTP for secure group messaging, we created an AI-driven incident command room. Inside each group thread, multiple agents have defined roles: a Remediation Agent identifies the issue and suggests fixes, an Audit Agent reviews the solution for safety, and a Documentation Agent generates a postmortem summary. Human engineers can monitor, approve, or intervene at any time. The group setting is essential to the project. Agents collaborate visibly in the same thread, maintaining shared context and structured role-based communication. This demonstrates real human-agent teamwork rather than a single chatbot response. Our tech stack includes Convos for multi-agent conversation handling, XMTP for decentralized messaging, Node.js and Express for backend logic, and React for the real-time chat interface. ClawOps showcases how group-based AI coordination can transform incident response into a collaborative, supervised, and intelligent workflow.