Lightweight, async-native job queue for Python

Documentation: https://chicory-dev.github.io/chicory

Source Code: https://github.com/chicory-dev/chicory

Chicory is a modern, async-first task queue for Python that makes distributed task execution simple and reliable. Built from the ground up with Python's async/await syntax, Chicory provides a clean, type-safe API for managing background jobs.

• Async Native - Built with Python's asyncio from day one
• Type Safe - Full type hints with automatic input/output validation via Pydantic
• Multiple Brokers - Support for Redis and RabbitMQ message brokers
• Multiple Backends - Store results in Redis, PostgreSQL, MySQL, SQLite, or MS SQL Server
• Flexible Delivery - Choose between at-least-once and at-most-once delivery guarantees
• Smart Retries - Built-in retry policies with exponential backoff
• Simple CLI - Easy worker management with the included CLI tool
• Production Ready - Designed for reliability and performance at scale

Python 3.11+

Chicory is built on these excellent libraries:

• Pydantic for data validation
• Typer for CLI (optional)

=== "Basic Installation"

```console
$ pip install chicory
```

=== "With Redis"

```console
$ pip install chicory[redis]
```

=== "With RabbitMQ"

```console
$ pip install chicory[rabbitmq]
```

=== "With PostgreSQL"

```console
$ pip install chicory[postgres]
```

=== "With CLI"

```console
$ pip install chicory[cli]
```

=== "Everything"

```console
$ pip install chicory[all]
```

Create a file tasks.py:

from chicory import Chicory, BrokerType, BackendType

# Initialize Chicory with Redis broker and backend
app = Chicory(
    broker=BrokerType.REDIS,
    backend=BackendType.REDIS,
)

@app.task()
async def send_email(to: str, subject: str, body: str) -> dict:
    """Send an email asynchronously."""
    # Your email sending logic here
    await asyncio.sleep(1)  # Simulate work
    return {
        "status": "sent",
        "to": to,
        "subject": subject,
    }

@app.task()
async def process_data(data: dict) -> dict:
    """Process some data."""
    # Your data processing logic
    result = {
        "processed": True,
        "items": len(data.get("items", [])),
    }
    return result

Run workers to process tasks:

$ chicory worker tasks:app

Or programmatically:

from chicory import Worker

async def main():
    worker = Worker(app)
    await worker.start()

if __name__ == "__main__":
    import asyncio
    asyncio.run(main())

From your application code:

import asyncio
from tasks import app, send_email, process_data

async def main():
    # Connect to broker and backend
    await app.connect()

    try:
        # Dispatch a task and wait for result
        result = await send_email.delay(
            to="user@example.com",
            subject="Welcome!",
            body="Thanks for signing up!",
        )

        # Get the result (waits until task completes)
        email_result = await result.get(timeout=30.0)
        print(f"Email sent: {email_result}")

        # Fire-and-forget (don't wait for result)
        task_id = await process_data.send({"items": [1, 2, 3]})
        print(f"Task dispatched: {task_id}")

    finally:
        await app.disconnect()

if __name__ == "__main__":
    asyncio.run(main())

That's it! You now have a fully functional distributed task queue.

Chicory embraces Python's async capabilities from the start. No callbacks, no threading complexity—just clean async code:

@app.task()
async def fetch_user_data(user_id: int) -> dict:
    async with httpx.AsyncClient() as client:
        response = await client.get(f"https://api.example.com/users/{user_id}")
        return response.json()

# Dispatch and await
result = await fetch_user_data.delay(123)
data = await result.get()

Automatic validation of inputs and outputs using Pydantic:

from pydantic import BaseModel, EmailStr

class EmailData(BaseModel):
    to: EmailStr
    subject: str
    body: str

@app.task()
async def send_email(data: EmailData) -> dict:
    # data is automatically validated
    # Invalid input raises ValidationError before task is queued
    ...

# This validates the input immediately
await send_email.delay(EmailData(
    to="user@example.com",
    subject="Hello",
    body="World"
))

Choose the right tools for your needs:

• Brokers: Redis or RabbitMQ
• Backends: Redis, PostgreSQL, MySQL, SQLite, MS SQL Server
• Delivery Modes: At-least-once or at-most-once
• Validation Modes: Strict, inputs-only, outputs-only, or disabled

Scale from a single process to hundreds of workers across multiple machines.

Built-in support for:

• Automatic retries with exponential backoff
• Task context for accessing task metadata
• Worker lifecycle hooks for setup/teardown
• Dead-letter queues for failed tasks
• Task timeouts and cancellation
• Graceful shutdown with configurable drain periods

Chicory is not a drop-in replacement for Celery, but if you're starting a new async Python project, Chicory provides a more modern and Pythonic experience.

For detailed documentation, visit https://chicory-dev.github.io/chicory

• Tutorial: Step-by-step guides for common use cases
• User Guide: In-depth documentation of all features
• API Reference: Complete API documentation

Contributions are welcome!

This project is licensed under the MIT License - see the LICENSE file for details.

Chicory is built and maintained by:

• Davide Di Mauro (@dadodimauro)
• Andrea Vouk (@andreavou)

Inspired by the excellent work of the Celery and FastAPI projects.