Examples Overview

Work through these examples in order to progressively learn PydanticAI — each one builds on concepts from the previous.

Learning Path

• 1. Direct Model Requests

  ---

  Understand basic LLM API calls without agents.

  Beginner

• 2. Temperature

  ---

  Control model output randomness and creativity.

  Beginner

• 3. Reasoning Effort

  ---

  Tune reasoning depth for speed vs accuracy trade-offs.

  Beginner

• 4. Basic Sentiment

  ---

  Learn structured outputs with Pydantic models and Literal types.

  Intermediate

• 5. Dynamic Classification

  ---

  Generate classification schemas at runtime with create_model().

  Intermediate

• 6. Bielik (Local Models)

  ---

  Run local models with Ollama, use tools and multi-turn conversations.

  Intermediate

• 7. History Processor

  ---

  Manage conversation history, filtering, persistence, and context windows.

  Advanced

• 8. OCR Parsing

  ---

  Extract structured data from PDFs with schema validation and async processing.

  Advanced