Pydantic-AI Deepagents (opens in new tab)

Python deep agent framework built on top of Pydantic AI designed to help you quickly build production-grade autonomous agents with planning, filesystem operations, subagent delegation, and skills.

A complete demonstration of all pydantic-deep features in a web application. -> (demo repository) https://github.com/vstorm-co/pydantic-deep/blob/main/example... -> (demo video) https://drive.google.com/file/d/1hqgXkbAgUrsKOWpfWdF48cqaxRh...

Features

    Multiple Backends: StateBackend (in-memory), FilesystemBackend, DockerSandbox, CompositeBackend
    Rich Toolsets: TodoToolset, FilesystemToolset, SubAgentToolset, SkillsToolset
    File Uploads: Upload files for agent processing with run_with_files() or deps.upload_file()
    Skills System: Extensible skill definitions with markdown prompts
    Structured Output: Type-safe responses with Pydantic models via output_type
    Context Management: Automatic conversation summarization for long sessions
    Human-in-the-Loop: Built-in support for human confirmation workflows
    Streaming: Full streaming support for agent responses

## Installation

```bash pip install pydantic-deep ```

Or with uv:

```bash uv add pydantic-deep ```

### Optional dependencies

```bash # Docker sandbox support pip install pydantic-deep[sandbox] ```

## Quick Start

```python import asyncio from pydantic_deep import create_deep_agent, create_default_deps from pydantic_deep.backends import StateBackend

async def main(): # Create a deep agent with state backend backend = StateBackend() deps = create_default_deps(backend) agent = create_deep_agent()

    # Run the agent
    result = await agent.run("Help me organize my tasks", deps=deps)
    print(result.output)

asyncio.run(main()) ```

## Structured Output

Get type-safe responses with Pydantic models:

```python from pydantic import BaseModel from pydantic_deep import create_deep_agent, create_default_deps

class TaskAnalysis(BaseModel): summary: str priority: str estimated_hours: float

agent = create_deep_agent(output_type=TaskAnalysis) deps = create_default_deps()

result = await agent.run("Analyze this task: implement user auth", deps=deps) print(result.output.priority) # Type-safe access ```

## File Uploads

Process user-uploaded files with the agent:

```python from pydantic_deep import create_deep_agent, DeepAgentDeps, run_with_files from pydantic_deep.backends import StateBackend

agent = create_deep_agent() deps = DeepAgentDeps(backend=StateBackend())

# Upload and process files with open("sales.csv", "rb") as f: result = await run_with_files( agent, "Analyze this sales data and find top products", deps, files=[("sales.csv", f.read())], ) ```

Or upload files directly to deps:

```python deps.upload_file("config.json", b'{"key": "value"}') # File is now at /uploads/config.json and agent sees it in system prompt ```

## Context Management

Automatically summarize long conversations to manage token limits:

```python from pydantic_deep import create_deep_agent from pydantic_deep.processors import create_summarization_processor

processor = create_summarization_processor( trigger=("tokens", 100000), # Summarize when reaching 100k tokens keep=("messages", 20), # Keep last 20 messages )

agent = create_deep_agent(history_processors=[processor]) ```

## Documentation

- *[Full Documentation](https://vstorm-co.github.io/pydantic-deep/)* - Complete guides and API reference - *[PyPI Package](https://pypi.org/project/pydantic-deep/)* - Package information and releases - *[GitHub Repository](https://github.com/vstorm-co/pydantic-deep)* - Source code and issues

### Quick Links

- [Installation Guide](https://vstorm-co.github.io/pydantic-deep/installation/) - [Core Concepts](https://vstorm-co.github.io/pydantic-deep/concepts/) - [Examples](https://vstorm-co.github.io/pydantic-deep/examples/) - [API Reference](https://vstorm-co.github.io/pydantic-deep/api/)