Observability platform that records agent sessions, visualizes execution paths, and tracks LLM costs for debugging production workflows.

MCP server that lets AI assistants inspect recorded traces and spans to diagnose agent failures directly from their development environment.

Open-source command-line tool that scans repository documentation for spelling, grammar, and code-example issues using language models.

Open-source Python and TypeScript instrumentation libraries that capture model calls, tools, traces, and agent activity with minimal application changes.

Open-source Python library that estimates prompt and completion costs using token counts and model pricing data.

Open-source AI agent that generates reusable jq queries to transform arbitrary JSON inputs into a requested schema.