pydantic-ai-todo 0.1.2

Todo/task planning toolset for pydantic-ai agents

pydantic-ai-todo

Looking for a complete agent framework? Check out pydantic-deep - a full-featured deep agent framework with planning, subagents, and skills system built on pydantic-ai.

Need file storage or Docker sandbox? Check out pydantic-ai-backend - file storage and sandbox backends that work with any pydantic-ai agent.

Want a full-stack template? Check out fastapi-fullstack - production-ready AI/LLM application template with FastAPI, Next.js, and pydantic-deep integration.

Todo/task planning toolset for pydantic-ai agents.

This library was extracted from pydantic-deep to provide standalone task planning for any pydantic-ai agent without requiring the full framework.

Provides read_todos and write_todos tools that help AI agents track and manage tasks during a session.

Installation

pip install pydantic-ai-todo

Or with uv:

uv add pydantic-ai-todo

Quick Start

from pydantic_ai import Agent
from pydantic_ai_todo import create_todo_toolset

# Create an agent with todo capabilities
agent = Agent(
    "openai:gpt-4.1",
    toolsets=[create_todo_toolset()],
)

# Run the agent
result = await agent.run("Create a todo list for building a website")

Usage with Storage Access

If you need to access the todos after the agent runs:

from pydantic_ai import Agent
from pydantic_ai_todo import create_todo_toolset, TodoStorage

# Create storage and toolset
storage = TodoStorage()
toolset = create_todo_toolset(storage=storage)

agent = Agent("openai:gpt-4.1", toolsets=[toolset])

# Run the agent
result = await agent.run("Plan the implementation of a REST API")

# Access todos directly
for todo in storage.todos:
    print(f"[{todo.status}] {todo.content}")

Custom Storage

You can implement custom storage (e.g., for persistence):

import json
from pydantic_ai_todo import create_todo_toolset, TodoStorageProtocol, Todo

class RedisTodoStorage:
    """Store todos in Redis."""

    def __init__(self, redis_client):
        self._redis = redis_client

    @property
    def todos(self) -> list[Todo]:
        data = self._redis.get("todos")
        if not data:
            return []
        return [Todo(**t) for t in json.loads(data)]

    @todos.setter
    def todos(self, value: list[Todo]) -> None:
        self._redis.set("todos", json.dumps([t.model_dump() for t in value]))

# Use with agent
storage = RedisTodoStorage(redis.Redis())
agent = Agent("openai:gpt-4.1", toolsets=[create_todo_toolset(storage)])

System Prompt Integration

Include current todos in the system prompt:

from pydantic_ai_todo import get_todo_system_prompt, TodoStorage

storage = TodoStorage()
# ... agent populates todos ...

# Generate system prompt section with current todos
prompt_section = get_todo_system_prompt(storage)

API Reference

create_todo_toolset(storage=None, *, id=None)

Creates a todo toolset with read_todos and write_todos tools.

Parameters:

• storage: Optional TodoStorageProtocol implementation. Defaults to in-memory TodoStorage.
• id: Optional unique ID for the toolset.

Returns: FunctionToolset[Any]

get_todo_system_prompt(storage=None)

Generates a system prompt section for task management.

Parameters:

• storage: Optional storage to read current todos from.

Returns: str - System prompt with optional current todos section.

Todo

Pydantic model for a todo item.

class Todo(BaseModel):
    content: str  # Task description in imperative form
    status: Literal["pending", "in_progress", "completed"]
    active_form: str  # Present continuous form (e.g., "Implementing...")

TodoStorage

Default in-memory storage implementation.

storage = TodoStorage()
storage.todos = [Todo(...), Todo(...)]
print(storage.todos)

TodoStorageProtocol

Protocol for custom storage implementations. Must have a todos property with getter and setter.

Related Projects

• pydantic-ai - The foundation: Agent framework by Pydantic
• pydantic-deep - Full agent framework (uses this library)
• pydantic-ai-backend - File storage and sandbox backends
• fastapi-fullstack - Full-stack AI app template

License

MIT