Skip to content

taskito

Rust-powered task queue for Python. No broker required — just SQLite.

pip install taskito

5-Minute Quickstart

from taskito import Queue

queue = Queue(db_path="tasks.db")

@queue.task()
def add(a: int, b: int) -> int:
    return a + b

job = add.delay(2, 3)

# Start worker (in production, use the CLI instead)
import threading
t = threading.Thread(target=queue.run_worker, daemon=True)
t.start()

print(job.result(timeout=10))  # 5

Get started

Why taskito?

Most Python task queues require a separate broker (Redis, RabbitMQ) that you need to install, configure, monitor, and keep running. taskito embeds everything into a single SQLite file — the queue, the results, the rate limits, the schedules. Just pip install and go.

The core engine is written in Rust for performance: job dispatch, retry scheduling, rate limiting, and storage all happen in compiled native code. Python only runs during actual task execution.

Features

  • Zero Infrastructure

    No Redis, no RabbitMQ — just a SQLite file. Install and start queuing in seconds.

  • Priority Queues

    Higher priority jobs run first. Override at enqueue time for urgent work.

  • Retry with Backoff

    Automatic exponential backoff with jitter. Failed jobs land in a dead letter queue for inspection.

  • Rate Limiting

    Token bucket rate limiting per task. "100/m", "10/s", "3600/h".

  • Task Workflows

    Compose pipelines with chain, parallelize with group, aggregate with chord.

  • Cron Scheduling

    @queue.periodic(cron="0 */5 * * * *") for recurring tasks with 6-field cron expressions.

  • Progress Tracking

    Report progress from inside tasks. Monitor completion percentage in real time.

  • Rust-Powered

    Scheduling, storage, and dispatch in native Rust. Python only runs your task code.

Architecture

graph TB
    subgraph Python ["Python Layer"]
        A["Queue / TaskWrapper"]
        D["JobResult"]
    end

    subgraph Rust ["Rust Core · PyO3"]
        F["PyQueue"]
        G["Scheduler · Tokio"]
        H["Worker Pool · OS Threads"]
        I["Rate Limiter"]
    end

    subgraph Storage ["Embedded Storage"]
        J[("SQLite · WAL mode<br/>Diesel ORM")]
    end

    A -->|enqueue| F
    F -->|INSERT| J
    G -->|poll & dequeue| J
    G -->|crossbeam channel| H
    H -->|acquire GIL · run task| A
    H -->|result| G
    G -->|UPDATE status| J
    D -->|poll status| F
    F -->|SELECT| J
    G -.->|check limit| I
    I -.->|token state| J

Architecture deep dive

Comparison

Feature taskito Celery RQ Dramatiq Huey
Broker required No Redis/RabbitMQ Redis Redis/RabbitMQ Redis
Core language Rust + Python Python Python Python Python
Priority queues ✅ ✅ ❌ ❌ ✅
Rate limiting ✅ ✅ ❌ ✅ ❌
Dead letter queue ✅ ❌ ✅ ❌ ❌
Task workflows ✅ ✅ ❌ ✅ ❌
Job cancellation ✅ ✅ ❌ ❌ ✅
Progress tracking ✅ ✅ ❌ ❌ ❌
Unique tasks ✅ ❌ ❌ ❌ ✅
Setup complexity pip install Broker + backend Redis server Broker Redis server

Full comparison