Favicon of Agent Console

Agent Console

A local AI coding agent dashboard tracks Claude Code and Codex tokens, cache use and estimated costs, with a self-hosted hub for multiple machines.

Agent Console helps developers and teams monitor Claude Code and Codex sessions through a local dashboard. It reads the agents' transcripts on your machines so you can compare usage and spot sessions that need attention.

  • Token and cost tracking: Compare usage over time, by model and by machine, with separate totals for cache reads, cache writes, uncached input and output. Costs are estimates based on published model rates; missing readings appear as unknown.
  • Session and cache analysis: See session activity, project and branch, model, observed subagents and context size, then inspect input growth and possible cache breaks.
  • Local alerts: Flag repeated identical tool calls, unusually large token responses and spending without an observed successful tool result. These signals don't prove an agent is stuck, and the alert panel covers only the current machine.
  • Shared team view: Connect laptops, build machines and servers through a self-hosted hub to compare their usage in one place without counting the same usage twice. Presenting mode replaces project, machine and person names with stand-ins.
  • Monitoring integrations: Optionally accept Claude Code OpenTelemetry and Kong or LiteLLM token metrics, or export metrics to Prometheus and Grafana. Telemetry readings stay separate from transcript totals to avoid duplicate counts.

The console and reporting clients run on macOS, Linux and Windows. It's open source under the MIT license, with Apple-signed and notarized macOS builds. It sends no telemetry, update checks or crash reports.

Similar to Agent Console