AI Agent Tracing Quickstart

Give your autonomous engineer eyes to see your AI. Choose between automatic setup with the CLI or manual integration with decorators.

Two setup options: CLI automatically generates code, or manual setup where you add decorators. For complete autonomous engineer setup, use our Main Quickstart.

Option 1: CLI Setup (Recommended)

Automatic code generation for your agent functions:

terminal


npm install -g @handit.ai/cli
handit-cli setup

The CLI detects your agent functions and generates tracing code automatically. No manual coding required.

Setup complete! Your autonomous engineer is now monitoring your AI automatically.

Option 2: Manual Setup

Add tracing decorators to your main agent functions yourself:

Install and configure

Add tracing to your agent

✅ DO trace (agent entry points):

❌ DON’T trace helper functions:


# These get captured automatically - don't add @tracing here
async def classify_intent(message):
async def search_knowledge(query):
async def generate_response(context):

Key point: Only trace your main agent functions that handle user requests. Everything inside gets captured automatically.

Verify setup

Go to dashboard.handit.ai - you should see agent traces appearing when you run your AI.

Next steps

Set Up Evaluation: Add Quality Assessment so your autonomous engineer can detect quality issues.

Enable Autonomous Fixes: Connect GitHub Integration so your autonomous engineer can create pull requests with improvements.

Your autonomous engineer can now see everything! Add evaluation and GitHub integration to complete your setup.