Takk

Define your architecture in pure Python-servers, workers, scheduled jobs, and databases connect automatically through type hints.

Overview

Takk is a Python framework that eliminates configuration overhead by using type hints to automatically wire up your infrastructure. Write pure Python code and let Takk handle the connections between servers, workers, databases, and scheduled jobs.

No YAML. No configuration files. Just Python.

Features

• Type-hint driven - Your type annotations define your architecture
• Automatic dependency injection - Components connect without manual wiring
• Pure Python - Everything is code, nothing is configuration
• Full IDE support - Autocomplete and type checking work out of the box
• Minimal boilerplate - Focus on your logic, not setup

Installation

uv add takk

Quick Start

from takk.models import Project, FastAPIApp, Worker
from takk.secrets import SlackWebhook

from my_app.settings import AppSettings
from my_app import app

background_worker = Worker("background")

project = Project(
    name="my-custom-server",
    shared_settings=[AppSettings],

    background_worker=background_worker,

    my_server=FastAPIApp(app),
)

How It Works

Takk uses Python type hints to understand your application's resources and automatically creates the necessary connections. When you annotate a settings class with a type like PostgresDsnor RedisDsn, Takk:

1. Detects the dependency through type
2. Instantiates the component with appropriate configuration
3. Injects it into your environment

Read the full article to see how we built this approach.

Core Components

Server

from takk.models import Project, FastAPIApp, Worker
from takk.secrets import SlackWebhook

project = Project(
    name="my-custom-server",

    custom_network_app=NetworkApp(
        command=["/bin/bash", "-c", "uv run main.py"],
        port=8000,
    ),
)

Worker

from takk import Worker

worker = Worker("name-of-worker")

worker.run(function, Args(...))

Database

from pydantic import PostgresDsn, RedisDsn, BaseModel
from takk import Database

class MyAppSettings(BaseModel):
    redis_url: RedisDsn
    psql_db: PostgresDsn

Scheduled Jobs

from takk.models import Compute, Project, Job

from my_app.train import train_model, TrainConfig

project = Project(
    name="ml-example",

    train_pokemon_model=Job(
        train_model,
        cron_schedule="0 3 * * *",  # Runs daily at 3 AM
        arguments=TrainConfig(...),
    ),
)

Requirements

• Python 3.10+
• Type hints support

Examples

Check out the examples directory for complete applications:

• Simple web server
• Background worker with scheduling
• Full-stack application

Learn More

• Documentation