Phoenix
Open Source
Monitoring
Open-source AI observability for experimentation and evaluation
About
Phoenix is Arize AI's open-source observability platform designed for experimentation, evaluation, and troubleshooting of AI applications. Accepts traces via standard OpenTelemetry Protocol (OTLP) with auto-instrumentation for LlamaIndex, LangChain, DSPy, Vercel AI SDK, OpenAI, Anthropic, and Bedrock. Features trace scoring with LLM-based evaluators or code checks, datasets and experiments for version comparison, and deployment flexibility (local, Jupyter, containers, cloud). Used by PepsiCo, Handshake, Tripadvisor, and Siemens. Fully open-source with 7,800+ GitHub stars and no feature gates.
Compatibility
Supported Languages
python
typescript
java
Details
- Category
- Monitoring