Structured Responses: Python vs YAML

This example demonstrates two ways to define structured responses in LLMling-agent:

• Using Python Pydantic models
• Using YAML response definitions
• Type validation and constraints
• Agent integration with structured outputs

How It Works

1. Python-defined Responses:

2. Use Pydantic models

3. Full IDE support and type checking
4. Best for programmatic use

5. Inline field documentation

6. YAML-defined Responses:

7. Define in configuration

8. Include validation constraints
9. Best for configuration-driven workflows
10. Self-documenting fields

Example Output:

Python-defined Response:
Main point: User expresses enthusiasm for new feature
Is positive: true

YAML-defined Response:
Sentiment: positive
Confidence: 0.95
Mood: excited

This demonstrates:

• Two ways to define structured outputs
• Validation and constraints
• Integration with type system
• Trade-offs between approaches

Open in Pydantic Playground