Show HN: Unified local observability for AI coding assistants (github.com)

AI Observer has launched a self-hosted, OpenTelemetry-compatible observability backend tailored for monitoring local AI coding tools such as Claude Code, Gemini CLI, and OpenAI Codex CLI. This innovative tool offers developers a unified dashboard to track critical metrics like token usage, costs, API latency, and error rates in real-time, enhancing visibility and cost management of AI tools without relying on external services. With features like customizable widgets and a built-in data analytics engine powered by DuckDB, users can efficiently analyze large datasets and maintain historical data for better debugging. The significance of AI Observer lies in its capacity to streamline the monitoring process for increasingly essential AI coding assistants, addressing a major gap in understanding their behaviors and expenses. By keeping telemetry data local, developers can ensure privacy and comply with regulations while gaining insights that can directly inform development strategies. The tool's multi-architecture Docker support also facilitates deployment across various environments, making it a flexible solution for organizations eager to integrate AI tools into their workflows while maintaining control and oversight.