Logfire
Pydantic's OpenTelemetry-native observability for Python.
About
Logfire is Pydantic's observability platform: OpenTelemetry-native logging, tracing, and metrics with the DX you'd expect from the pydantic team — structured logs, auto-instrumentation for popular libraries (FastAPI, httpx, OpenAI...), and a polished dashboard. 4.5k stars, MIT, 13M PyPI downloads/month. It's general-purpose observability, not LLM-specific — pair it with Langfuse/Opik for prompt-level tracing. Note: the hosted platform is the real product; the OSS library shines brightest with a Logfire account.
What it can do
- **OTel-native**: standard OpenTelemetry under the hood
- **Auto-instrumentation**: FastAPI, httpx, OpenAI, databases
- **Structured logging**: f-string style with rich context
- **Tracing + metrics**: full observability triad
- **Pydantic DX**: the polish you'd expect
Setup
```bash
pip install logfire
```
```python
import logfire
logfire.configure() # or logfire.auth()
logfire.info("Hello, {name}!", name="world")
```
Auto-instruments FastAPI, httpx, SQLAlchemy, OpenAI, and more. Docs: logfire.pydantic.dev.
Prerequisites
Python 3.9+. Logfire account for the hosted dashboard.
Use cases
Works with
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: 4,513 GitHub stars (log-scaled vs catalog peers); blended with ~13,028,487 monthly package downloads; last push 2d ago; 47 releases/12mo; 210 open issues. Manual review: Pydantic's OTel-native observability, 4.5k stars, 13M PyPI/mo; hosted platform is the real product; general observability, not LLM-specific. License: MIT.
★ 4,513 GitHub stars
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 7.9/10
- Maintenance · recency, releases, issue load
- 8.2/10
- Openness · license + docs
- 10.0/10
- Manual review · sanity read + red-flag check
- 8.0/10
- · 4,513 GitHub stars (log-scaled vs catalog peers)
- · blended with ~13,028,487 monthly package downloads
- · last push 2d ago; 47 releases/12mo; 210 open issues
- · manual review: Pydantic's OTel-native observability, 4.5k stars, 13M PyPI/mo; hosted platform is the real product; general observability, not LLM-specific
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.