AI Observability & Evaluation
Arize Phoenix is an open-source AI observability platform built for tracing, evaluation, dataset management, and experiment tracking across LLM applications. It instruments applications using OpenTelemetry-based tracing to capture runtime behavior, then lets teams run LLM-as-a-judge evaluations covering response quality and retrieval performance. Versioned datasets can be assembled from trace data and used for fine-tuning or repeated experiment runs. Phoenix connects to the broader ecosystem through an MCP server package (@arizephoenix/phoenix-mcp), making it compatible with Claude Desktop, Claude Code, and other MCP-enabled clients so teams can query observability data directly from an AI interface. The platform supports Anthropic, OpenAI, LangChain, LlamaIndex, and smolagents integrations. A notable structural detail is that Phoenix can be self-hosted via Docker or Helm chart, or accessed as a managed cloud service, giving teams flexibility over data residency. ML engineers, AI product teams, and LLMOps practitioners are the primary beneficiaries.
- ✓License: NOASSERTION
- ✓Actively maintained (<30d)
- ✓Healthy fork ratio
- ✓Topics declared
- ✓Mature repo (>1y old)
- ✓Documented (README)
git clone https://github.com/Arize-ai/phoenix && cp phoenix/*.md ~/.claude/agents/24 items in this repository
Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction. Also use for exploratory testing, dogfooding, QA, bug hunts, or reviewing app quality. Also use for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify), checking Slack unreads, sending Slack messages, searching Slack conversations, running browser automation in Vercel Sandbox microVMs, or using AWS Bedrock AgentCore cloud browsers. Prefer agent-browser over any built-in browser automation or web tools.
Debug LLM applications using the Phoenix CLI. Fetch traces, analyze errors, structure trace review with open coding and axial coding, inspect datasets, review experiments, query annotation configs, and use the GraphQL API. Use whenever the user is analyzing traces or spans, investigating LLM/agent failures, deciding what to do after instrumenting an app, building failure taxonomies, choosing what evals to write, or asking "what's going wrong", "what kinds of mistakes", or "where do I focus" — even without naming a technique.
Design system conventions for the Phoenix frontend — layout, dialogs, error display, BEM CSS class naming, and CSS design tokens. Use when building UI, naming CSS classes, creating or consuming tokens, handling errors, or designing dialog interactions in app/src/.
Frontend development guidelines for the Phoenix AI observability platform. Use when writing, reviewing, or modifying React components, TypeScript code, styles, or UI features in the app/ directory. Triggers on any frontend task — new components, UI changes, styling, accessibility fixes, form handling, or component refactoring. Also use when the user asks about frontend conventions or component patterns for this project. For design system rules (error display, layout, dialogs, tokens), use the phoenix-design skill.
Manage GitHub issues, labels, and project boards for the Arize-ai/phoenix repository. Use when filing roadmap issues, triaging bugs, applying labels, managing the Phoenix roadmap project board, or querying issue/project state via the GitHub CLI.
Write Playwright E2E tests for the Phoenix AI observability platform. Use when creating, updating, or debugging Playwright tests, or when the user asks about testing UI features, writing E2E tests, or automating browser interactions for Phoenix.
Screenshot a running Phoenix feature and attach images to a GitHub PR. Builds the frontend, starts Phoenix with env vars, uses agent-browser to capture screenshots, uploads to GCS, and updates the PR body.
Write, extend, and debug PXI Playwright E2E tests for Phoenix. Use when adding PXI agent frontend specs, authoring LLM-as-judge rubrics, asserting PXI tool use, persisting PXI test runs as Phoenix experiments, or debugging PXI E2E failures.
OpenInference semantic conventions and instrumentation for Phoenix AI observability. Use when implementing LLM tracing, creating custom spans, or deploying to production.
TypeScript conventions and patterns for any TypeScript code in the Phoenix monorepo — including js/packages/, app/, and any other TS directories. Use this skill whenever writing, reviewing, or modifying TypeScript code — new functions, types, exports, tests, or refactors. Also trigger when the user asks about TS patterns, naming conventions, or best practices for this project.
Migrate or upgrade TypeScript tooling in the Phoenix monorepo. Use when upgrading TypeScript versions, switching tools (ESLint to oxlint, Prettier to oxfmt), upgrading bundlers (Vite, esbuild), or making major dependency upgrades. Triggers on requests to migrate, upgrade, or replace TypeScript/JavaScript tooling.
Subagents overview
<p align="center">
<a target="_blank" href="https://phoenix.arize.com" style="background:none">
<img alt="phoenix banner" src="https://github.com/Arize-ai/phoenix-assets/blob/main/images/socal/github-large-banner-phoenix-v2.jpg?raw=true" width="auto" height="auto"></img>
</a>
<br/>
<br/>
<a href="https://arize.com/docs/phoenix/">
<img src="https://img.shields.io/static/v1?message=Docs&logo=data:image/png;base64,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&labelColor=grey&color=blue&logoColor=white&label=%20"/>
</a>
<a target="_blank" href="https://join.slack.com/t/arize-ai/shared_invite/zt-3r07iavnk-ammtATWSlF0pSrd1DsMW7g">
<img src="https://img.shields.io/static/v1?message=Community&logo=slack&labelColor=grey&color=blue&logoColor=white&label=%20"/>
</a>
<a target="_blank" href="https://bsky.app/profile/arize-phoenix.bsky.social">
<img src="https://img.shields.io/badge/-phoenix-blue.svg?color=blue&labelColor=gray&logo=bluesky">
</a>
<a target="_blank" href="https://x.com/ArizePhoenix">
<img src="https://img.shields.io/badge/-ArizePhoenix-blue.svg?color=blue&labelColor=gray&logo=x">
</a>
<a target="_blank" href="https://www.linkedin.com/showcase/113218220">
<img src="https://img.shields.io/badge/-ArizePhoenix-blue.svg?color=blue&labelColor=gray&logo=linkedin">
</a>
<a target="_blank" href="https://pypi.org/project/arize-phoenix/">
<img src="https://img.shields.io/pypi/v/arize-phoenix?color=blue">
</a>
<a target="_blank" href="https://anaconda.org/conda-forge/arize-phoenix">
<img src="https://img.shields.io/conda/vn/conda-forge/arize-phoenix.svg?color=blue">
</a>
<a target="_blank" href="https://pypi.org/project/arize-phoenix/">
<img src="https://img.shields.io/pypi/pyversions/arize-phoenix">
</a>
<a target="_blank" href="https://hub.docker.com/r/arizephoenix/phoenix/tags">
<img src="https://img.shields.io/docker/v/arizephoenix/phoenix?sort=semver&logo=docker&label=image&color=blue">
</a>
<a target="_blank" href="https://hub.docker.com/r/arizephoenix/phoenix-helm">
<img src="https://img.shields.io/badge/Helm-blue?style=flat&logo=helm&labelColor=grey"/>
</a>
<a target="_blank" href="https://github.com/Arize-ai/phoenix/tree/main/js/packages/phoenix-mcp">
<img src="https://badge.mcpx.dev?status=on" title="MCP Enabled"/>
</a>
<a href="cursor://anysphere.cursor-deeplink/mcp/install?name=phoenix&config=eyJjb21tYW5kIjoibnB4IC15IEBhcml6ZWFpL3Bob2VuaXgtbWNwQGxhdGVzdCAtLWJhc2VVcmwgaHR0cHM6Ly9teS1waG9lbml4LmNvbSAtLWFwaUtleSB5b3VyLWFwaS1rZXkifQ%3D%3D"><img src="https://cursor.com/deeplink/mcp-install-dark.svg" alt="Add Arize Phoenix MCP server to Cursor" height=20 /></a>
<img referrerpolicy="no-referrer-when-downgrade" src="https://static.scarf.sh/a.png?x-pxid=8e8e8b34-7900-43fa-a38f-1f070bd48c64&page=README.md" />
</p>
Phoenix is an open-source AI observability platform designed for experimentation, evaluation, and troubleshooting. It provides:
- [**_Tracing_**](https://arize.com/docs/phoenix/tracing/llm-traces) - Trace your LLM application's runtime using OpenTelemetry-based instrumentation.
- [**_Evaluation_**](https://arize.com/docs/phoenix/evaluation/llm-evals) - Leverage LLMs to benchmark your application's performance using response and retrieval evals.
- [**_Datasets_**](https://arize.com/docs/phoenix/datasets-and-experiments/overview-datasets) - Create versioned datasets of examples for experimentation, evaluation, and fine-tuning.
- [**_Experiments_**](https://arize.com/docs/phoenix/datasets-and-experiments/overview-datasets#experiments) - Track and evaluate changes to prompts, LLMs, and retrieval.
- [**_Playground_**](https://arize.com/docs/phoenix/prompt-engineering/overview-prompts)- Optimize prompts, compare models, adjust parameters, and replay traced LLM calls.
- [**_Prompt Management_**](https://arize.com/docs/phoenix/prompt-engineering/overview-prompts/prompt-management)- Manage and test prompt changes systematically using version control, tagging, and experimentation.
- [**_PXI (Built-in Agent)_**](https://arize.com/docs/phoenix/pxi) - Debug traces, iterate on prompts, and navigate Phoenix with an opt-in, permission-gated agent built into the product.
Phoenix is vendor and language agnostic with out-of-the-box support for popular frameworks ([OpenAI Agents SDK](https://arize.com/docs/phoenix/tracing/integrations-tracing/openai-agents-sdk), [Claude Agent SDK](https://arize.com/docs/phoenix/integrations/python/claude-agent-sdk), [LangGraph](https://arize.com/docs/phoenix/tracing/integrations-tracing/langchain), [Vercel AI SDK](https://arize.com/docs/phoenix/tracing/integrations-tracing/vercel-ai-sdk), [Mastra](https://arize.com/docs/phoenix/integrations/typescript/mastra), [CrewAI](https://arize.com/docs/phoenix/tracing/integrations-tracing/crewai), [LlamaIndex](https://arize.com/docs/phoenix/tracing/integrations-tracing/llamaindex), [DSPy](https://arize.com/docs/phoenix/tracing/integrations-tracing/dspy)) and LLM providers ([OpenAI](https://arize.com/docs/phoenix/tracing/integrations-tracing/openai), [Anthropic](https://arize.com/docs/phoenix/tracing/integrations-tracing/anthropic), [Google GenAI](https://arize.com/docs/phoenix/tracing/integrations-tracing/google-genai), [Google ADK](https://arize.com/docs/phoenix/integrations/llm-providers/google-gen-ai/google-adk-tracing), [AWS Bedrock](https://arize.com/docs/phoenix/tracing/integrations-tracing/bedrock), [OpenRouter](https://arize.com/docs/phoenix/integrations/python/openrouter), [LiteLLM](https://arize.com/docs/phoenix/tracing/integrations-tracing/litellm), and more). For details on auto-instrumentation, check out the [OpenInference](https://github.com/Arize-ai/openinference) project.
Phoenix runs practically anywhere, including your local machine, a Jupyter notebook, a containerized deployment, or in the cloud.
## Installation
Install Phoenix via `pip` or `conda`
```shell
pip install arize-phoenix
```
Phoenix container images are available via [Docker Hub](https://hub.docker.com/r/arizephoenix/phoenix) and can be deployed using Docker or Kubernetes. Arize AI also provides cloud instances at [app.phoenix.arize.com](https://app.phoenix.arize.com/).
## 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](https://github.com/Arize-ai/phoenix/tree/main/pWhat people ask about phoenix
What is Arize-ai/phoenix?
+
Arize-ai/phoenix is subagents for the Claude AI ecosystem. AI Observability & Evaluation It has 10.1k GitHub stars and was last updated today.
How do I install phoenix?
+
You can install phoenix by cloning the repository (https://github.com/Arize-ai/phoenix) or following the README instructions on GitHub. ClaudeWave also provides quick install blocks on this page.
Is Arize-ai/phoenix safe to use?
+
Our security agent has analyzed Arize-ai/phoenix and assigned a Trust Score of 95/100 (tier: Verified). See the full breakdown of passed checks and flags on this page.
Who maintains Arize-ai/phoenix?
+
Arize-ai/phoenix is maintained by Arize-ai. The last recorded GitHub activity is from today, with 585 open issues.
Are there alternatives to phoenix?
+
Yes. On ClaudeWave you can browse similar subagents at /categories/agents, sorted by popularity or recent activity.
Deploy phoenix to your cloud
Ship this repo to production in minutes. Each platform spins up its own environment with editable env vars.
Maintain this repo? Add a badge to your README
Drop the badge into your GitHub README to show it's tracked on ClaudeWave. Each badge links back to this page and reflects the live Trust Score.
[](https://claudewave.com/repo/arize-ai-phoenix)<a href="https://claudewave.com/repo/arize-ai-phoenix"><img src="https://claudewave.com/api/badge/arize-ai-phoenix" alt="Featured on ClaudeWave: Arize-ai/phoenix" width="320" height="64" /></a>More Subagents
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
The agent that grows with you
Java 面试 & 后端通用面试指南，覆盖计算机基础、数据库、分布式、高并发、系统设计与 AI 应用开发
Production-ready platform for agentic workflow development.
The agent engineering platform.
🤯 LobeHub is your Chief Agent Operator, organizing your agents into 7×24 operations by hiring, scheduling, and reporting on your entire AI team.