🧪 Clawmetry: Observability for AI Agents

The direction here is clear: Clawmetry aims to provide zero-configuration observability and governance for AI agents. It acts as a dashboard that lets you see exactly what your autonomous tools are doing behind the scenes. It supports tracking for 26 different AI runtimes—including Cursor, Claude Code, and Aider—monitoring live token costs, session histories, and tool calls. The interesting part is how it addresses a growing blind spot in the developer workflow. As more engineers rely on autonomous coding agents and local LLM tools, tracking API spend and understanding which local tools an agent actually invoked becomes critical. Clawmetry aggregates this telemetry into a single view. The hard part is not simply logging data, but maintaining compatibility across a fragmented ecosystem of agent frameworks that all handle execution differently. It explores a practical approach to AI governance, offering a much-needed layer of transparency for teams and power users managing multiple agentic tools.