AIエージェント
openllmetry
OpenTelemetry-native observability for LLM apps.
★ 7.5k スターApache-2.0
ガイドを見る →logfireは、pydanticが公開しているオープンソースのAIエージェントプロジェクトです（GitHub: pydantic/logfire）。GitHubスターは4,510、フォークは309で、MITライセンスの下で公開されています。GitHubでの最終更新は2026年10月8日です。
Pydantic Logfire: Python and AI observability on OpenTelemetry — SQL-queryable traces, metrics, and logs
Use the Logfire cloud platform (free tier available) or export OpenTelemetry data to another backend.
Install the SDK from PyPI and run the auth command if you will send data to Logfire.
Call logfire.configure() in your application entry point.
Add auto-instrumentation (for example for FastAPI) and view your first trace as a structured, queryable timeline.
Built on OTel, so data can also flow to any OTel-compatible backend — no proprietary agent.
Query traces, metrics, and logs with plain SQL instead of a bespoke query language.
Data is shown as rich Python objects rather than flattened JSON strings.
Instrument frameworks such as FastAPI with calls like logfire.instrument_fastapi(app).
Purpose-built views and evals for LLM calls and agent behavior.
See how your Pydantic models validate in production, from the team behind Pydantic itself.
Clone the project and enter its directory (most users install from PyPI instead).
git clone https://github.com/pydantic/logfire.git
cd logfireInstall the SDK and authenticate with the Logfire platform.
pip install logfire
logfire authAdd configuration and auto-instrumentation to your app, then run it and open the Logfire UI.
import logfire
logfire.configure()
logfire.instrument_fastapi(app)Pydantic — the team behind the Pydantic data-validation library and Pydantic AI.
The Python SDK in this repo is MIT-licensed and free. The hosted platform has free and paid tiers per the README, its backend is closed source, and self-hosting the platform requires an enterprise license.
No paid key is required to start: the SDK is free, the cloud has a free tier, and you can export to any OpenTelemetry backend instead.
Logfire is OpenTelemetry-native and SQL-queryable, spans general Python observability (not only LLMs), and renders data as Python objects; LangSmith focuses on LLM tracing within its own platform.
Run `pip install logfire`, `logfire auth`, add `logfire.configure()` to your app, and instrument your framework.
OpenTelemetry-native observability for LLM apps.
Python SDK for Agent AI Observability, Monitoring and Evaluation Framework. Includes features like agent, llm and tools tracing, debugging multi-agentic syst…
Supercharge LLM app evaluations with objective metrics, synthetic test data and feedback loops.
Use LLMs as plain Python functions: prompts become typed, testable code.
Nocturne_memory, in Python — built for a lightweight, rollbackable, and visual long-ter...
Modular Python framework for AI agents and workflows with tools and memory.
LibreAIカタログに掲載されたAIエージェントプロジェクトの中で最もスターの多いもの — logfireが合わなかった場合に検討したい、現実的な候補リストです。
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
カタログの中でlogfireとスター数が近いAIエージェントプロジェクト — 前後の棚もそのまま見て回れます。
M-flow is a bio-inspired cognitive memory engine for AI agents — a graph alternative to flat RAG
Agent framework for the JVM. Pronounced Em-BAY-bel /ɛmˈbeɪbəl/
LongMemory: local-first memory for AI agents with temporal truth, provenance, and hybrid recall
Memory-LanceDB-Pro gives OpenClaw agents hybrid search, smart extraction, and intelligent forgetting
Convene 18 legendary thinkers as an AI council that debates your hardest questions — inside Claude Code
The Ruby-native AI framework. Chats, agents, tools, images, audio, and video through one consistent API, in plain Ruby or Rails.