OpenLLMetry
OpenLLMetry is an open-source observability tool built on OpenTelemetry for LLM applications. It provides complete tracing of LLM calls, vector DBs, and frameworks, and can connect to existing observability solutions. It is for developers who want to monitor and debug their LLM applications.
✨ Key features
- Built on OpenTelemetry, connects to existing observability stacks
- Supports many LLM providers, vector DBs, and frameworks
- Easy SDK setup with Traceloop.init()
- Standard OpenTelemetry data output
- No telemetry collected in SDK v0.49.2+
- Apache 2.0 license
🎯 Use cases
- Trace LLM API calls to OpenAI, Anthropic, etc.
- Monitor vector database operations in Chroma, Pinecone, etc.
- Instrument LangChain, LlamaIndex, and other frameworks
- Send traces to Datadog, Honeycomb, or other backends
- Debug and optimize LLM application performance
📦 Installation
🧰 Requirements: Python environment with pip; no API keys required for basic setup, but you need an account for your chosen observability backend.
pip install traceloop-sdk
🚀 Usage
from traceloop.sdk import Traceloop
Traceloop.init()
To see traces immediately when running locally, disable batch sending:
Traceloop.init(disable_batch=True)
❓ FAQ
What is OpenLLMetry?
OpenLLMetry is a set of extensions built on OpenTelemetry that gives you complete observability over your LLM application.
How do I get started?
Install the SDK with pip install traceloop-sdk, then add Traceloop.init() to your code.
Can I use my existing observability tools?
Yes, because it uses OpenTelemetry, it can connect to Datadog, Honeycomb, and many other solutions.
Does OpenLLMetry collect telemetry?
No, the SDK and instrumentations do not log or collect any telemetry as of v0.49.2 and above.
📊 Repository
🤖 Overview, features, install steps and FAQ were generated from the project's README on Sep 4, 2026. Always check the original source before running commands.