Project Information
Category
- Inference
- Orchestration
- Runtime
- Training
Lookspan is a local-first observability dashboard for AI agents — MCP-native and OpenTelemetry-based. It captures the spans your agents emit and visualizes them locally (waterfall timeline, trace comparison), with cost tracking, replay + LLM-as-judge evaluation, dataset experiments, and CSV/JSON/HTML export. SQLite by default or Postgres. Ships drop-in SDK wrappers (@lookspan/openai, @lookspan/anthropic), an MCP integration (@lookspan/mcp), and Python SDKs. MIT licensed.
Project Information
Category
Lookspan is a local-first observability dashboard for AI agents — MCP-native and OpenTelemetry-based. It captures the spans your agents emit and visualizes them locally (waterfall timeline, trace comparison), with cost tracking, replay + LLM-as-judge evaluation, dataset experiments, and CSV/JSON/HTML export. SQLite by default or Postgres. Ships drop-in SDK wrappers (
@lookspan/openai,@lookspan/anthropic), an MCP integration (@lookspan/mcp), and Python SDKs. MIT licensed.