---
id: arize-phoenix-otel
version: "0.17.1"
license: Apache-2.0
license_treatment: permissive
maintenance: active
---

# arize-phoenix-otel — LLM Observability

License: permissive · Maintenance: active · Downloads: 2.0M/mo

## What it is and what it does

arize-phoenix-otel is a lightweight wrapper around OpenTelemetry that simplifies tracing setup for AI applications by providing Phoenix-aware defaults, automatic environment variable configuration, and drop-in replacements for standard OTel classes. It reads configuration from environment variables (PHOENIX_COLLECTOR_ENDPOINT, PHOENIX_API_KEY, PHOENIX_PROJECT_NAME, etc.) or a `.env.phoenix` file, with explicit arguments taking precedence.

The package's main value is reducing boilerplate: a single `register()` call with `auto_instrument=True` automatically instruments popular AI frameworks (OpenAI, LangChain, LlamaIndex) with zero code changes, batches spans for production use, and sends traces to a Phoenix instance over OTLP using OpenInference semantic conventions. It also provides tracing decorators for GenAI patterns and works with any recent Phoenix server version without version pairing.

Use it for:

- Enable automatic tracing of OpenAI, LangChain, or LlamaIndex calls in an LLM application with a single function call and environment variables.
- Collect and export distributed traces from a multi-component AI system to a self-hosted or cloud Phoenix instance for debugging and monitoring.
- Manually instrument custom AI workflows using decorators like @tracer.chain and @tracer.tool for GenAI-specific patterns.
- Configure production-ready batching and authentication for trace export without writing custom OpenTelemetry span processor code.
- Organize traces by project and environment using Phoenix project names and custom headers for multi-tenant or multi-environment deployments.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Wraps OpenTelemetry with Phoenix-aware defaults and decorators to instrument AI applications for tracing, with automatic configuration from environment variables and support for GenAI patterns.

Yes. Active maintenance, no vulnerabilities, low install friction, and permissive license make it a safe choice. Install it if you need to trace AI applications and are already using or planning to adopt Phoenix for observability; the auto-instrumentation and environment-based configuration save significant boilerplate compared to raw OpenTelemetry setup.

## Install

pip install arize-phoenix-otel
uv add arize-phoenix-otel
poetry add arize-phoenix-otel

## Installing arize-phoenix-otel

Before you install: Low friction: pure Python wheel with 8 runtime dependencies, all from the OpenTelemetry ecosystem. Active maintenance (last commit 2026-08-14, release 4 days old) and no known vulnerabilities.

License in practice: Apache-2.0 permissive license allows use in most commercial and open-source projects without significant restriction.

Quickstart:

pip install arize-phoenix-otel

from phoenix.otel import register

tracer_provider = register(
    auto_instrument=True,
    project_name="my-app",
    endpoint="http://localhost:6006/v1/traces"
)

Requires Python 3.10 or later. Auto-instrumentation of specific AI libraries (OpenAI, LangChain, LlamaIndex) requires installing corresponding openinference-instrumentation-* packages separately.

Verify before relying:

- Whether auto_instrument=True works without explicit instrumentation library imports or if they must be installed first
- Performance overhead of automatic instrumentation on typical AI application workloads
- Backward compatibility guarantees across Phoenix server versions

## Package facts

- License: Apache-2.0 (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 2.0M/month (top 5,000 on PyPI)
- Known vulnerabilities: none known

## Tags

opentelemetry tracing for llm applications, phoenix otel instrumentation, automatic ai library tracing, llm observability and monitoring, distributed tracing for generative ai, openinference semantic conventions, trace collection and export, llm-observability, opentelemetry, distributed-tracing

[View on SkillFed](https://skillfed.io/packages/arize-phoenix-otel) · [View on PyPI](https://pypi.org/project/arize-phoenix-otel/)