Helicone
Open-source LLM observability through a proxy or SDK
Helicone logs LLM requests by sitting in front of the provider API, or through an SDK. It tracks cost, latency and errors per user and prompt, and the code is open source.
More in Observability
Arize Phoenix
Observability
AI observability and evaluation you can self-host
Braintrust
Observability
Evals, logging and prompt tooling for AI products
Datadog LLM Observability
Observability
LLM traces inside Datadog
Langfuse
Observability
Open-source LLM observability, evals and prompt management
LangSmith
Observability
Tracing, evals and deployment from the LangChain team
OpenLLMetry
Observability
OpenTelemetry instrumentation for LLM apps