Arize Phoenix
Open-source LLM observability with ML monitoring roots
Our Take
Phoenix brings Arize's ML monitoring expertise to the LLM space. The OpenTelemetry-based instrumentation is a standout — it means you're not locked into a proprietary tracing format. Particularly strong for RAG evaluation. Phoenix 2.0 added a full web UI with dashboards, making it viable for platform teams beyond just notebook-based exploration.
Pros
- + OpenTelemetry-native (no vendor lock-in)
- + Strong RAG evaluation tools
- + Backed by established ML monitoring company
- + Fully open source
Cons
- - Web UI still catching up to notebook experience
- - Smaller community than Langfuse
Details
- Pricing Model
- open-source
- Starting Price
- $0
- Self-Hosted
- Yes
- Cloud Hosted
- Yes
- Founded
- 2020
- Repository
- GitHub →
Best For
- • LLM tracing with OpenTelemetry
- • Retrieval evaluation (RAG)
- • Teams already using Arize for ML monitoring
- • Dashboard-based and notebook exploration
Integrations
OpenAI LangChain LlamaIndex OpenTelemetry Anthropic CrewAI Vercel AI SDK
Articles featuring Arize Phoenix
Last updated: