Introduction

Sazabi is an AI-native observability platform. You connect your logs, and the Sazabi agent helps you find and fix production issues by answering questions about your systems — with the evidence it found — or by detecting issues and bringing you the answer before you ask.

If you want to start doing rather than reading, go to the Quickstart. This page explains what Sazabi is; the rest of this section gets you set up.

Philosophy

Sazabi is built on a few opinions: observability should do the investigative work for you, a conversation is a better interface than a dashboard, and everything is a log. See Philosophy for the beliefs behind how the product works.

What Sazabi is

Sazabi replaces the dashboard-and-query workflow of traditional observability with a conversation. You send your logs to Sazabi, and then you either:

• Ask. Open a thread and describe a problem in plain language. The Sazabi agent runs the queries, reads the results, and answers with the specific log lines and correlations it found.
• Get told. Sazabi watches your data and opens an issue when it detects a problem worth tracking, then notifies your team through the channels you configure.

Everything Sazabi does is grounded in your logs. It does not guess; it searches, and it shows the queries it ran so you can check its work.

How it works

1. Connect your data. Send logs through OpenTelemetry, a native SDK, or a direct connection to a platform such as AWS CloudWatch, Google Cloud, or Vercel. See Instrument.
2. Get alerted, or ask. Sazabi detects issues and notifies you, and you can open a thread to ask a question at any time.
3. Get answers with evidence. The agent replies with the log lines, patterns, and correlations it found, and shows every query it ran.
4. Use it from anywhere. The web dashboard, Slack, Microsoft Teams, your coding agent, the CLI, and the API all reach the same threads and data.

In this section