---
id: traceloop-sdk
version: "0.62.3"
license: Apache-2.0
license_treatment: permissive
maintenance: active
---

# traceloop-sdk — Traceloop Software Development Kit (SDK) for Python

License: permissive · Maintenance: active · Downloads: 2.5M/mo

## What it is and what it does

Traceloop SDK is a Python instrumentation library that captures execution traces from LLM applications—calls to OpenAI, Anthropic, LangChain, LlamaIndex, and other frameworks—and exports them to Traceloop's cloud platform or to your own OpenTelemetry-compatible observability stack. It wraps your LLM code non-intrusively, meaning you initialize it once and then your existing application code generates traces automatically without modification.

The SDK includes built-in instrumentation for many LLM providers and frameworks (OpenAI, Anthropic, Bedrock, Cohere, Groq, LangChain, LlamaIndex, CrewAI, and others), plus vector databases and logging. You choose where traces go—Traceloop's managed service or your own observability backend via OTLP gRPC or HTTP exporters. It's designed for debugging, monitoring, and understanding LLM application behavior in development and production.

Use it for:

- Debug LLM application failures by inspecting full execution traces including all API calls and responses.
- Monitor token usage and latency across multiple LLM providers in a single dashboard.
- Export traces to an existing observability platform via OpenTelemetry OTLP exporters.
- Track performance of LangChain or LlamaIndex workflows across complex multi-step operations.
- Audit and log all interactions with external LLM APIs for compliance or cost analysis.

## Worth the install?

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

Traceloop SDK instruments LLM applications to capture execution traces and send them to Traceloop or your own observability backend, built on OpenTelemetry.

Yes, if you are building or operating LLM applications and need visibility into their behavior. The SDK is actively maintained, has no known vulnerabilities, and integrates seamlessly with popular frameworks. Install friction is low and the permissive license poses no barrier. Verify whether Traceloop's managed service fits your cost model and whether the instrumentation overhead is acceptable for your workload.

## Install

pip install traceloop-sdk
uv add traceloop-sdk
poetry add traceloop-sdk

## Installing traceloop-sdk

Before you install: Low friction install as a pure Python wheel. Active maintenance with a recent release (4 days old) and strong community signal (7377 stars). 50 runtime dependencies, mostly OpenTelemetry instrumentation packages for specific LLM frameworks, means you pay the cost of what you actually use.

License in practice: Apache-2.0 permissive license allows commercial and private use with minimal restrictions.

Quickstart:

pip install traceloop-sdk

from traceloop.sdk import Traceloop

Traceloop.init(app_name="my_app")
# Your LLM code is now traced

Requires Python 3.10 or later.

Verify before relying:

- Whether traces sent to Traceloop incur costs or require a paid account.
- Performance overhead of tracing in production workloads.
- How to configure which LLM framework instrumentations are loaded selectively.
- Full list of supported LLM providers and frameworks beyond those in runtime dependencies.

## Package facts

- License: Apache-2.0 (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 2.5M/month (top 5,000 on PyPI)
- Known vulnerabilities: none known

## Tags

llm tracing and monitoring, opentelemetry llm instrumentation, ai application observability, llm debugging and traces, openai call tracing, distributed tracing for ai, langchain observability, llm-observability, opentelemetry, tracing

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