absurd-sdk

What it is and what it does

Absurd SDK provides a Python interface to Absurd, a durable execution system built entirely on PostgreSQL. It lets you define long-running workflows as task functions that survive crashes, restarts, and network failures by decomposing work into recorded steps. Each step's result is persisted, so if a worker crashes mid-execution, the system can resume from the last completed step without duplicating work.

The SDK offers both synchronous and asynchronous APIs. You register tasks as decorated functions, define steps within them using either decorators (sync) or async/await patterns (async), and optionally wait for external events before proceeding. The system handles scheduling, retries, and state management automatically, requiring only a Postgres connection—no additional services, brokers, or infrastructure.

Use it for:

• Order fulfillment workflows that process payments, reserve inventory, await shipment events, and send notifications across multiple steps.
• Multi-step data pipelines where intermediate results must survive worker restarts and be resumable from the last completed stage.
• Scheduled background jobs with built-in retry logic and durability, replacing simpler queue systems when crash-safety is required.
• Event-driven workflows that pause execution to await external events (e.g., webhook callbacks) before resuming downstream steps.
• Parent-child task hierarchies where one task spawns and durably waits for results from child tasks without losing state.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Python SDK for Absurd, a PostgreSQL-based durable task execution system that handles scheduling, retries, and long-lived workflows without requiring additional services beyond Postgres.

Yes. The package is actively maintained, has low install friction, carries no known vulnerabilities, and solves a real problem—durable task execution without external infrastructure—in a way that's genuinely simpler than multi-service alternatives if you already run Postgres. The alpha status reflects early maturity, but the recent release cadence and community interest (2335 stars) suggest it's production-ready for teams comfortable with that trade-off.

Install

absurd-sdk on PyPI

pip install absurd-sdk

uv add absurd-sdk

poetry add absurd-sdk

Installing absurd-sdk

Before you install

Low install friction with a single runtime dependency (psycopg). Active maintenance with a recent release (10 days ago) and 2335 repository stars indicate ongoing development and community interest.

License in practice

Licensed under Apache-2.0 (permissive), allowing use in commercial and private projects with minimal restrictions.

Quickstart

pip install absurd-sdk

from absurd_sdk import Absurd

app = Absurd("postgresql://localhost/absurd")

@app.register_task(name="my-task")
def my_task(params, ctx):
    @ctx.run_step("step-name")
    def step_work():
        return {"result": 42}
    return step_work

app.start_worker()

Requires a running PostgreSQL database; connection URL can be set via ABSURD_DATABASE_URL, PGDATABASE environment variables, or passed explicitly.

Verify before relying

• Whether Postgres schema is automatically initialized on first connection or requires manual setup
• Performance characteristics and scalability limits for high-volume task workloads
• Retry policy configuration options and backoff strategies
• Support for task priorities or custom scheduling logic

Package facts

Evidence: absurd_sdk-0.5.0-py3-none-any.whl

Keywords: absurd, postgresql, queue, task, durable, workflow

Tags

More Application Frameworks packages