MerchantryTidbits

evals-testing / cli

langfuse

Capability: langfuse

Use it when

  • observe LLM traces in production
  • trace prompts costs scores for agent runs
  • eval agent quality observability
  • self-host LLM tracing
  • production prompt and trajectory observability you can host yourself

What it solves

Not the fit when

  • replaces product analytics

Install

Self-host or cloud from https://github.com/langfuse/langfuse

Invoke

Trace prompts, costs, scores for agent runs when babysitting false completes.

Alternatives

No reviewed alternatives recorded yet.