Release Notes v0.2.1

Release Notes v0.2.1

🎉 New Feature: Pydantic Support

chak now natively supports Pydantic models for tool parameters and return values, bringing type safety and automatic validation to your LLM tool calls.

✨ What's New

• Type-Safe Function Parameters: Use Pydantic models as function parameters - automatic JSON → object conversion
• Type-Safe Return Values: Return Pydantic models from tools - automatic object → JSON serialization
• Stateful Objects + Pydantic: Combine state persistence with type safety for complex business logic
• Zero Configuration: Just use Pydantic type hints - no decorators or wrappers needed

📝 Example

from pydantic import BaseModel, Field

class UserInput(BaseModel):
    name: str = Field(description="User's full name")
    email: str = Field(description="Email address")
    age: int

class UserOutput(BaseModel):
    id: int
    name: str
    status: str = "active"

def create_user(user: UserInput) -> UserOutput:
    """Create a new user"""
    return UserOutput(id=123, name=user.name)

conv = chak.Conversation(
    "openai/gpt-4o",
    tools=[create_user]
)

# LLM automatically validates and converts!
response = await conv.asend("Create user: John Doe, john@example.com, 30")