Arize Phoenix

Arize Phoenix provides an open‑source platform for tracing, evaluating, and optimizing large language models (LLMs). It automatically instruments LLM calls, captures request and response data, and stores the information for later analysis. The system includes a prompt playground where users can experiment with different prompts and view the resulting model behavior in real time.

The platform offers real‑time monitoring of AI applications, presenting metrics such as latency, token usage, and error rates through dashboards. It supports integration with Python and Conda environments, is available as a Docker image, and can be deployed via Helm charts. Documentation, community channels, and versioned releases are linked from the project homepage.

Designed for developers and data scientists who need visibility into LLM performance, Phoenix helps identify issues, compare model variants, and guide optimization efforts without requiring manual instrumentation of each model call.