Arize Phoenix
arize.com
- Free plan
- Web
- API
- Self-hosted
What Arize Phoenix is
Arize Phoenix helps teams trace, evaluate, debug, and improve AI applications. It records model calls, retrieval, tool use, and custom logic in traces that can be sent over OpenTelemetry. Teams can score traces and spans using LLM-based evaluators, code checks, or human labels, then rerun datasets across application versions to compare evaluation results. Prompt tools provide versioning, side-by-side tests, and call replay to inspect how changes affect outputs. Documented integrations include LlamaIndex, LangChain, DSPy, OpenAI, Bedrock, and Anthropic, alongside evaluation libraries such as Harbor, Ragas, Cleanlab, and MLflow. Phoenix can run locally or be self-hosted with Docker, Kubernetes, or a cloud environment. Arize says self-hosting keeps traces in the user’s own environment. Phoenix is free and open source under the ELv2 license, with users managing trace spans, ingestion volume, projects, and retention. Self-hosted deployments support authentication and OAuth2 identity providers with an OpenID Connect well-known configuration endpoint.
Who it suits
Phoenix suits teams that need to inspect and evaluate AI application traces, compare application versions, or manage prompts. Its local and self-hosted options may suit teams that want traces kept in their own environment.
Good to know, and what to watch
- Captures model calls, retrieval, tools, and custom logic.
- Supports evaluator, code, and human scoring.
- Includes prompt versioning, testing, and call replay.
- Can run locally or be self-hosted.
- Users manage trace spans, ingestion, projects, and retention.
- Self-hosted OAuth2 requires an OpenID Connect well-known endpoint.
Plans
Every plan and pricePhoenix
Free plan
The verdict
Phoenix combines trace inspection with evaluation, prompt tools, and dataset-based comparisons. Teams should plan how to manage usage and retention, particularly when self-hosting.
What the maker publishes
- Authentication
- Self-hosted Phoenix supports authentication and OAuth2 identity providers that expose an OpenID Connect well-known configuration endpoint. source
- Data control
- Arize says self-hosting keeps traces in the user’s own environment. source
- Deployment
- Phoenix can run locally or be self-hosted using Docker, Kubernetes, or a cloud environment. source
- Evaluation integrations
- Phoenix can use evaluation libraries including Harbor, Ragas, Cleanlab, and MLflow. source
- Evaluations
- Teams can score traces and spans with LLM-based evaluators, code checks, or human labels. source
- Experiments
- Teams can rerun datasets through different application versions and compare evaluation results. source
- Integrations
- Documented integrations include frameworks and providers such as LlamaIndex, LangChain, DSPy, OpenAI, Bedrock, and Anthropic. source
- License
- The Phoenix page describes the project as ELv2 licensed. source
- Maker
- Phoenix is built by Arize AI and the open-source community. source
- Prompt tools
- Phoenix supports prompt versioning, side-by-side testing, and replaying calls to inspect how changes affect outputs. source
- Purpose
- Phoenix helps teams trace, evaluate, debug, and improve AI applications. source
- Support
- The Phoenix documentation points users to a Slack community for questions and developer discussion. source
- Tracing
- It captures model calls, retrieval, tool use, and custom logic in traces that can be sent over OpenTelemetry. source
- Usage limits
- Arize describes Phoenix as free and open source, with user-managed trace spans, ingestion volume, projects, and retention. source
Checked against the maker's pages on 5 Oct 2026.