Chat with Your Data

DeepTracer's AI chat lets you ask questions about your production data in plain English. No query languages, no dashboard hunting — just ask what you want to know.

How it works

The AI has access to tools that query your logs, errors, traces, and health data. When you ask a question, it decides which data to pull, runs the queries, and gives you a clear answer.

Example prompts

Here are some things you can ask:

What errors happened in the last hour?

Show me the most common errors this week

Why is the API slow?

Find all payment-related errors

How many users hit the 500 error on /api/checkout today?

What changed since yesterday? Any new errors?

The AI adapts to your question. It might search logs, pull up error groups, check trace latencies, or combine data from multiple sources to give you a complete answer.

What the AI can do

Under the hood, the AI uses these capabilities:

• Search logs — Filter by level, service, time range, and free-text
• Query errors — Look up error groups, occurrence counts, and trends
• Check traces — Find slow requests and analyze latency
• Read health data — Get an overview of your app's current status
• Cross-reference — Connect errors to the logs and traces around them

Chat history

Your conversations are saved per project. Come back later to review what you asked, or pick up where you left off. Each project maintains its own chat history, so context stays relevant.

Also available in your editor

DeepTracer chat is also available as an MCP (Model Context Protocol) integration. This means you can ask the same questions directly from:

• Cursor — Setup guide
• Claude Code — Setup guide
• Windsurf — Setup guide

Ask "what errors are happening in production?" without leaving your editor.

Plan limits

On the Free plan, each message you send counts toward your monthly limit. The AI's responses don't count against your limit.

What's next?