Show HN: Dispytch – Python framework for event-driven services

4 months ago 17

Dispytch is a lightweight, async-first Python framework for event-handling. It’s designed to streamline the development of clean and testable event-driven services.

🧠 Async-first core – built for modern Python I/O
🔌 FastAPI-style dependency injection – clean, decoupled handlers
📬 Backend-flexible – with Kafka and RabbitMQ out-of-the-box
🧱 Composable architecture – extend, override, or inject anything
🧾 Pydantic-based validation – event schemas are validated using pydantic
🔁 Built-in retry logic – configurable, resilient, no boilerplate

Install using uv with extras for your preferred backend:

for Kafka support:

For RabbitMQ support:

uv add dispytch[rabbitmq]

from typing import Annotated from pydantic import BaseModel from dispytch import Event, Dependency, HandlerGroup from service import UserService, get_user_service class User(BaseModel): id: str email: str name: str class UserCreatedEvent(BaseModel): user: User timestamp: int user_events = HandlerGroup() @user_events.handler(topic='user_events', event='user_registered') async def handle_user_registered( event: Event[UserCreatedEvent], user_service: Annotated[UserService, Dependency(get_user_service)] ): user = event.body.user timestamp = event.body.timestamp print(f"[User Registered] {user.id} - {user.email} at {timestamp}") await user_service.do_smth_with_the_user(event.body.user)

import uuid from datetime import datetime from pydantic import BaseModel from dispytch import EventBase class User(BaseModel): id: str email: str name: str class UserEvent(EventBase): __topic__ = "user_events" class UserRegistered(UserEvent): __event_type__ = "user_registered" user: User timestamp: int async def example_emit(emitter): await emitter.emit( UserRegistered( user=User( id=str(uuid.uuid4()), email="[email protected]", name="John Doe", ), timestamp=int(datetime.now().timestamp()), ) )

While dispytch is a great choice for most usecases there are some limitations to be aware of:

🧾 No schema-on-write support Dispytch uses a schema-on-read model. Formats like Avro, Protobuf, or Thrift aren’t supported yet.

🕵️ No dead-letter queue (DLQ) Failed messages are retried using built-in logic, but there’s no DLQ or fallback mechanism after final retries yet.

🧩 No topic pattern matching Wildcard or templated subscriptions (e.g. user.*, order:{id}:events) aren’t supported in handler declarations yet. Though the backend of you choice may route events to the appropriate queues/topics/channels there is no way for a dispytch handler to know it yet

💡 See something missing? Some features aren’t here yet—but with your help, they could be. Contributions welcome via PRs or discussions.

Read Entire Article

Show HN: Dispytch – Python framework for event-driven services

Related

Barnton, Edinburgh's Nuclear Bunker

Self play and autocurricula in the age of agents

FDA Strips Breast Cancer Warning from Menopause Hormone Ther...