mudrii/openclaw-dashboard

What it solves

OpenClaw Dashboard provides a centralized, local command center for managing and monitoring complex AI agent deployments. It eliminates the need to dig through log files, CLI commands, and multiple data sources to track gateway health, costs, model usage, and the status of scheduled cron jobs and sub-agents.

How it works

The project is a zero-dependency application consisting of a single Go binary backend and a pure HTML/CSS/JS frontend. It collects data from an OpenClaw installation via a CLI adapter and exposes it through a local web interface. The backend implements a "stale-while-revalidate" pattern for data refreshes, ensuring the UI remains responsive while updating the system state in the background.

Who it’s for

Users running OpenClaw AI agents at scale—specifically those managing multiple agents, numerous cron jobs, various communication channels (Telegram, WhatsApp, Slack, Discord), and multiple LLM models—who need an at-a-glance overview of their system's health and expenditures.

Highlights

• Comprehensive Observability: Live tracking of CPU/RAM, gateway status, and detailed per-model token usage and costs.
• Agent & Task Monitoring: Real-time visibility into active sessions, sub-agent activity, and cron job status (including "flapping" detection for unstable jobs).
• Integrated AI Chat: An optional chat interface that allows users to query their dashboard data using natural language via the OpenClaw gateway.
• Zero-Dependency Design: Runs as a local-only service on macOS and Linux with no external dependencies or cloud requirements.
• Customizable UI: Features a glass morphism interface with six built-in themes and support for custom CSS color variables.