OpenLLMetry is a Python-based OpenTelemetry extension suite for instrumenting GenAI/LLM apps, with a Traceloop SDK. It supports multiple integrations and destinations, and includes examples to initialize tracing via Traceloop.init().
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
OpenLLMetry is a set of extensions built on top of OpenTelemetry that provides observability for LLM applications. It includes standard OpenTelemetry instrumentations plus custom extensions for GenAI-related calls and vector DBs, and a Traceloop SDK to get started with OpenLLMetry.
How it works
The project builds on OpenTelemetry to emit traces from LLM-related operations and integrations. It offers instrumentations for various providers and vector databases, and a SDK that initializes tracing using Traceloop.init(), outputting standard OpenTelemetry data.
Getting started
The recommended start is via the SDK. Install the SDK:
pip install traceloop-sdk
Then instrument your code:
from traceloop.sdk import Traceloop
Traceloop.init()
If running locally and you want immediate traces, you can disable batch sending:
Traceloop.init(disable_batch=True)
Recent releases
Latest release is v0.62.1 (2026-06-28):
- Fix: bump version problem (#4332)
- Commit: bump: version 0.62.0 → 0.62.1
Previous: v0.62.0 (2026-06-28) with Feat: instrumentation: Add litellm package instrumentation (#4322)
Earlier: v0.61.0 (2026-05-31) with feats around exposing use_legacy_attributes and async support for bedrock/AIs etc.
Traction
Stars: 7353
Behind the repo
OpenLLMetry is maintained by Traceloop and is licensed under Apache-2.0. The repository shows a Python project with ongoing releases and a suite of integrations and instruments for OpenTelemetry-compatible observability.
Caveats
License: Apache-2.0 Created: 2023-09-02 Last push: 2026-08-04 Issues: 627 Languages: Python