Welcome to Neatlogs

Observability for AI agents: every run traced, every failure caught, every fix tracked.

Get started in one command

The wizard auto-detects your stack, installs the SDK, and instruments your app — no manual setup. Prefer to wire it up yourself? Manual setup.

Neatlogs is an observability platform for AI agents. Instrument your app in one line, then explore traces, catch failures, and collaborate on fixes from a single dashboard.

Instrument in one line — two lines of Python and every run (every LLM call, tool call, and retry) lands in your dashboard.

pip install -U neatlogs

import neatlogs
neatlogs.init()

Start here

Introduction

What Neatlogs does and how the pieces fit together.

Setup Wizard

Let the wizard detect your stack and instrument your code for you.

Explore the Dashboard

A guided tour of traces, detections, and Investigate.

Import Traces

Bring existing data over from Langfuse, LangSmith, Braintrust, or Raindrop.

Go deeper

Features

Traces, detections, Investigate, analytics, Fixes, experiments, and alerts.

Integrations

Setup guides for 25+ providers, frameworks, and developer tools.

Guides

End-to-end walkthroughs: RAG, multi-agent, chatbots, MCP.

SDK Reference

Python, TypeScript, and Browser SDKs, plus HTTP trace injection.

FAQ

Answers to common questions about Neatlogs.