zerodb-celery

Celery broker + result backend powered by ZeroDB. Replace Redis/RabbitMQ with one line.

Why?

Celery requires Redis or RabbitMQ for its broker and result backend. That means provisioning, configuring, and paying for infrastructure you shouldn't need.

zerodb-celery replaces both with ZeroDB — a cloud database that auto-provisions on first use. No signup, no credit card, no infrastructure.

Install

pip install zerodb-celery

Quick Start

from celery import Celery
from zerodb_celery import ZeroDBBroker, ZeroDBBackend

app = Celery('tasks')

# Configure both broker and backend
ZeroDBBroker.configure(app)
ZeroDBBackend.configure(app)

@app.task
def add(x, y):
    return x + y

# Trigger a task
result = add.delay(4, 6)
print(result.get(timeout=30))  # 10

Configuration

Auto-provisioning (default)

Just use zerodb://auto — a free ZeroDB project is created on first use:

app.config_from_object({
    'broker_url': 'zerodb://auto',
    'broker_transport': 'zerodb_celery.broker:Transport',
    'result_backend': 'zerodb_celery.backend:ZeroDBBackend',
})

Explicit credentials

Set environment variables:

export ZERODB_API_KEY=zdb_your_key_here
export ZERODB_PROJECT_ID=proj_your_id_here

Or pass them directly:

ZeroDBBroker.configure(app, api_key='zdb_...', project_id='proj_...')
ZeroDBBackend.configure(app, api_key='zdb_...', project_id='proj_...')

Credential resolution order

1. Explicit api_key / project_id parameters
2. ZERODB_API_KEY / ZERODB_PROJECT_ID environment variables
3. Cached credentials in ~/.zerodb/credentials.json
4. Auto-provision a new free project via API

How It Works

Broker (message queue)

The broker uses ZeroDB's event stream as a message queue:

• Publish task: POST /api/v1/zerodb/events with topic celery:{queue_name}
• Consume task: GET /api/v1/zerodb/events?topic=celery:{queue_name}&consume=true

Result backend (task results)

The backend stores results in a ZeroDB table called celery_results:

• Store result: POST /api/v1/zerodb/tables/celery_results/rows
• Get result: GET /api/v1/zerodb/tables/celery_results/rows/{task_id}

The table is auto-created on first use.

Use Cases

• ML pipelines: Queue training jobs, store results — no Redis needed
• AI agent workflows: Background task processing for agent swarms
• Serverless apps: No infrastructure to manage alongside your Celery workers
• Prototyping: Get Celery running in 30 seconds without Docker

Comparison

Development

# Install dev dependencies
pip install -e ".[dev]"

# Run tests
pytest tests/ -v

License

MIT