phoenix README

The project's README from github.com/Arize-ai/phoenix. As one markdown file: README.md.

phoenix banner

English · 简体中文

Add Arize Phoenix MCP server to Cursor

Arize Phoenix is Arize's open-source AI observability platform designed for experimentation, evaluation, and troubleshooting. For managed production workflows, Arize also offers Arize AX. Phoenix provides:

Phoenix is vendor and language agnostic with out-of-the-box support for popular frameworks (OpenAI Agents SDK, Claude Agent SDK, LangGraph, Vercel AI SDK, Mastra, CrewAI, LlamaIndex, DSPy) and LLM providers (OpenAI, Anthropic, Google GenAI, Google ADK, AWS Bedrock, OpenRouter, LiteLLM, and more). For details on auto-instrumentation, check out the OpenInference project.

Phoenix runs practically anywhere, including your local machine, a containerized deployment, or in the cloud. See Environments for a walkthrough of each option, or jump straight into the Tracing Quickstart.

[!NOTE] Phoenix is open-source and self-hosted. For a managed platform, Arize offers Arize AX — the same OpenTelemetry/OpenInference instrumentation works with both. See the full comparison.

Table of Contents

Run Locally

Install Phoenix via pip or conda and have a fully functional Phoenix. For all installation and hosting options, see the install guide.

pip install arize-phoenix
phoenix serve

Or run it with no install using uvx:

uvx arize-phoenix serve

Trace Your Application

The fastest way to send traces is to let your coding agent (Claude Code, Codex, Cursor, and others) instrument your app. From your project directory, run:

npx @arizeai/phoenix-cli setup
# or, with Phoenix installed: px setup

Setup detects your framework and LLM provider, installs the right OpenInference instrumentation, and wires up trace export. Prefer to wire it up in code? See the tracing documentation.

Deploy

Phoenix container images are available via Docker Hub and can be deployed using Docker or Kubernetes via the Helm chart.

For Docker Compose, Kubernetes/Helm, and other deployment options, see the self-hosting documentation.

Deploy on Railway   Deploy to Render   Run on Google Cloud   Deploy to Azure   Deploy to AWS

[!NOTE] The Google Cloud button builds Phoenix from source in Cloud Shell rather than deploying the prebuilt Docker Hub image. The Azure template serves plain HTTP (Azure Container Instances does not terminate TLS) — front it with a TLS proxy such as an Application Gateway before production use.

Packages

The arize-phoenix package includes the entire Phoenix platform. However, if you have deployed the Phoenix platform, there are lightweight Python sub-packages and TypeScript packages that can be used in conjunction with the platform.

Python Subpackages

Package Version & Docs Description
arize-phoenix-otel PyPI Version Docs Provides a lightweight wrapper around OpenTelemetry primitives with Phoenix-aware defaults
arize-phoenix-client PyPI Version Docs Lightweight client for interacting with the Phoenix server via its OpenAPI REST interface
arize-phoenix-evals PyPI Version Docs Tooling to evaluate LLM applications including RAG relevance, answer relevance, and more

TypeScript Subpackages

Package Version & Docs Description
@arizeai/phoenix-otel NPM Version Docs Provides a lightweight wrapper around OpenTelemetry primitives with Phoenix-aware defaults
@arizeai/phoenix-client NPM Version Docs Client for the Arize Phoenix API
@arizeai/phoenix-evals NPM Version Docs TypeScript evaluation library for LLM applications (alpha release)
@arizeai/phoenix-mcp NPM Version Docs Standalone stdio MCP server for older Phoenix versions (maintenance mode — superseded by the remote MCP server built into Phoenix)
@arizeai/phoenix-cli NPM Version Docs CLI for fetching traces, datasets, and experiments for use with Claude Code, Cursor, and other coding agents

Tracing Integrations

Phoenix is built on top of OpenTelemetry and is vendor, language, and framework agnostic. For details about tracing integrations and example applications, see the OpenInference project and the integrations documentation.

Python Integrations

Integration Package Version
OpenAI openinference-instrumentation-openai PyPI Version
OpenAI Agents openinference-instrumentation-openai-agents PyPI Version
LlamaIndex openinference-instrumentation-llama-index PyPI Version
DSPy openinference-instrumentation-dspy PyPI Version
AWS Bedrock openinference-instrumentation-bedrock PyPI Version
LangChain openinference-instrumentation-langchain PyPI Version
LangGraph openinference-instrumentation-langchain PyPI Version
MistralAI openinference-instrumentation-mistralai PyPI Version
Cohere openinference-instrumentation-cohere PyPI Version
Together AI openinference-instrumentation-together PyPI Version
Ollama openinference-instrumentation-ollama PyPI Version
Google GenAI openinference-instrumentation-google-genai PyPI Version
Google ADK openinference-instrumentation-google-adk PyPI Version
Guardrails openinference-instrumentation-guardrails PyPI Version
VertexAI openinference-instrumentation-vertexai PyPI Version
CrewAI openinference-instrumentation-crewai PyPI Version
Haystack openinference-instrumentation-haystack PyPI Version
LiteLLM openinference-instrumentation-litellm PyPI Version
OpenRouter openinference-instrumentation-openai PyPI Version
OrcaRouter openinference-instrumentation-openai PyPI Version
Groq openinference-instrumentation-groq PyPI Version
Instructor openinference-instrumentation-instructor PyPI Version
Anthropic openinference-instrumentation-anthropic PyPI Version
Smolagents openinference-instrumentation-smolagents PyPI Version
Agno openinference-instrumentation-agno PyPI Version
BeeAI openinference-instrumentation-beeai PyPI Version
Strands Agents openinference-instrumentation-strands-agents PyPI Version
Restate openinference-instrumentation-openai-agents PyPI Version
MCP openinference-instrumentation-mcp PyPI Version
Pydantic AI openinference-instrumentation-pydantic-ai PyPI Version
AG2 openinference-instrumentation-ag2 PyPI Version
Autogen AgentChat openinference-instrumentation-autogen-agentchat PyPI Version
Portkey openinference-instrumentation-portkey PyPI Version
Agent Spec openinference-instrumentation-agentspec PyPI Version
Claude Agent SDK openinference-instrumentation-claude-agent-sdk PyPI Version
TypeSafe AI openinference-instrumentation-typesafe PyPI Version

Span Processors

Normalize and convert data across other instrumentation libraries by adding span processors that unify data.

Package Description Version
openinference-instrumentation-openlit OpenInference Span Processor for OpenLIT traces. PyPI Version
openinference-instrumentation-openllmetry OpenInference Span Processor for OpenLLMetry (Traceloop) traces. PyPI Version

JavaScript Integrations

Integration Package Version
OpenAI @arizeai/openinference-instrumentation-openai NPM Version
OpenAI Agents @arizeai/openinference-instrumentation-openai-agents NPM Version
LangChain.js @arizeai/openinference-instrumentation-langchain NPM Version
TanStack AI @arizeai/openinference-tanstack-ai NPM Version
Vercel AI SDK @arizeai/openinference-vercel NPM Version
BeeAI @arizeai/openinference-instrumentation-beeai NPM Version
Claude Agent SDK @arizeai/openinference-instrumentation-claude-agent-sdk NPM Version
Mastra @mastra/arize NPM Version
MCP @arizeai/openinference-instrumentation-mcp NPM Version
TypeSafe AI @arizeai/openinference-instrumentation-typesafe NPM Version

Java Integrations

Integration Package Version
LangChain4j openinference-instrumentation-langchain4j Maven Central
SpringAI openinference-instrumentation-springAI Maven Central
Arconia for Spring AI io.arconia:arconia-openinference-semantic-conventions Maven Central
Google ADK openinference-instrumentation-adk-java Maven Central

Go Integrations

Integration Package Version
OpenAI github.com/Arize-ai/openinference/go/openinference-instrumentation-openai-go Go Reference
Anthropic github.com/Arize-ai/openinference/go/openinference-instrumentation-anthropic-sdk-go Go Reference

Platforms

Platform Description Docs
BeeAI AI agent framework with built-in observability Integration Guide
Dify Open-source LLM app development platform Integration Guide
Envoy AI Gateway AI Gateway built on Envoy Proxy for AI workloads Integration Guide
LangFlow Visual framework for building multi-agent and RAG applications Integration Guide
LiteLLM Proxy Proxy server for LLMs Integration Guide
Flowise Visual framework for building LLM applications Integration Guide
Prompt Flow Microsoft's prompt flow orchestration tool Integration Guide
NVIDIA NeMo NVIDIA NeMo Agent Toolkit for enterprise agents Integration Guide
Graphite Multi-agent LLM workflow framework with visual builder Integration Guide

Sandboxes

Run Phoenix code evaluators in hosted sandbox providers for kernel-level isolation and runtime dependency installation.

Integration Description Docs
E2B Hosted micro-VM sandboxes for AI-generated code Integration Guide
Daytona Managed development sandboxes with snapshot startup Integration Guide
Vercel Sandbox Ephemeral compute on Vercel's infrastructure Integration Guide
Modal Serverless, Python-first container platform Integration Guide

For Humans and Coding Agents

Phoenix is built to be driven by people and by AI coding agents alike. Three surfaces let agents (Claude Code, Codex, Cursor, and others) work with your traces, datasets, and experiments:

See the coding agents documentation for setup and usage.

Skill Description
phoenix-cli Debug LLM applications using the Phoenix CLI — fetch traces, analyze errors, review experiments, and query the GraphQL API
phoenix-evals Build and run evaluators for AI/LLM applications using Phoenix
phoenix-tracing OpenInference semantic conventions and instrumentation for tracing LLM applications

Security & Privacy

We take data security and privacy very seriously. For more details, see our Security and Privacy documentation.

Telemetry

By default, Phoenix collects basic web analytics (e.g., page views, UI interactions) to help us understand how Phoenix is used and improve the product. None of your trace data, evaluation results, or any sensitive information is ever collected.

You can opt-out of telemetry by setting the environment variable: PHOENIX_TELEMETRY_ENABLED=false

Community

Join our community to connect with thousands of AI builders.

Breaking Changes

See the migration guide for a list of breaking changes.

Copyright 2025 Arize AI, Inc. All Rights Reserved.

Portions of this code are patent protected by one or more U.S. Patents. See the IP_NOTICE.

This software is licensed under the terms of the Elastic License 2.0 (ELv2). See LICENSE.

phoenix (Arize-ai/phoenix)

Searching Database…

11.9K ·GitHub
Что это?Arize Phoenix — это платформа наблюдаемости для ИИ с открытым исходным кодом для трассировки, оценки и диагностики приложений на основе LLM, агентов и RAG-конвейеров.
Phoenix — это библиотека, которую я импортирую, или платформа, которую я запускаю?Это самостоятельно размещаемая платформа со своим сервером и интерфейсом для трассировок, наборов данных и экспериментов; помимо неё есть лёгкие подпакеты для Python и TypeScript (otel, client, evals, cli), которые можно использовать отдельно.
Как мне запустить Phoenix?Установите его через pip или conda либо запустите без установки через uvx; для продакшена можно развернуть готовый образ контейнера из Docker Hub с помощью Docker или Kubernetes/Helm, а затем поручить агенту для написания кода автоматически подключить инструментирование к вашему приложению из каталога проекта.
Что нужно, прежде чем попробовать?Python (для pip/conda) или Docker для развёртывания в контейнере; трассировка построена на инструментировании на основе OpenTelemetry, а Phoenix автоматически определяет ваш фреймворк (LangChain, LlamaIndex, CrewAI и др.) и провайдера LLM, чтобы всё это настроить.
Есть ли подводные камни при развёртывании, о которых стоит знать?Готовый шаблон для Azure работает только по обычному HTTP, потому что Azure Container Instances не завершает TLS, поэтому перед использованием в продакшене нужно поставить перед ним TLS-прокси (например, Application Gateway).
Что такое PXI (Phoenix Intelligence)?PXI — это ИИ-агент для инженерных задач, встроенный прямо в Phoenix; он помогает отлаживать трассировки, итеративно улучшать промпты и разбираться в продукте.
Могут ли агенты для программирования, такие как Claude Code или Cursor, работать с Phoenix напрямую?Да — Phoenix предоставляет удалённый сервер MCP по эндпоинту /mcp, чтобы агенты могли запрашивать трассировки, наборы данных и эксперименты, а также пакет npm phoenix-cli для получения этих данных из терминала.
С какими фреймворками и провайдерами LLM он работает из коробки?Для фреймворков из коробки есть инструментирование для OpenAI Agents SDK, Claude Agent SDK, LangGraph, LlamaIndex, CrewAI и DSPy, а для провайдеров — OpenAI, Anthropic, Google GenAI, AWS Bedrock и OpenRouter.
Есть ли управляемый вариант, если я не хочу размещать всё сам?Да, Arize предлагает Arize AX как управляемую платформу для продакшена, которая использует ту же инструментацию OpenTelemetry/OpenInference, что и открытый, самостоятельно размещаемый Phoenix.
Creator
Arize-ai
Size
538.8 MB
Stars
11.9K
Age
3y
Commits · 90 Days