Claude Code Agent Monitor

Claude Code Agent Monitor puts Claude Code, Cursor and Codex activity in one local dashboard for developers tracking coding sessions and teams monitoring agent work. It runs in a browser served by your own machine, with desktop apps for macOS and Windows and a VS Code sidebar. The dashboard works offline, requires no account and sends no telemetry. It's free and open source under the MIT license.

Live views show tool calls, agent status and conversations as work happens. A Kanban board separates work in progress from sessions waiting for input, completed tasks and errors. Searchable history and conversation replay help you inspect what an agent did, including its tool output and subagent activity. Cursor conversations remain available even after the editor removes its own history.

Token accounting handles context compression, while per-model and per-subagent cost estimates help you compare spending across sessions and projects. Workflow diagrams show delegation, tool transitions and collaboration between agents. Rules can flag stuck agents, error bursts or token thresholds and send alerts to webhook endpoints.

You can also start and resume Claude Code or Codex sessions in the browser, inspect their configuration and edit supported files with backups. Capture uses native hooks and local transcripts, so your CLI workflow can continue if the dashboard stops. It stores monitoring data locally in SQLite; optional SSH sources bring remote histories into the dashboard. A local MCP server, Prometheus metrics and Grafana dashboards support automation and team monitoring.