Simple concurrent runner for async tasks, threads and processes – one API, one object.

conrun is under active development. Until v1.0 the public API may change without notice. Feel free to experiment, but pin the exact version in production.

• One-liner concurrency – run any coroutine, function, or CPU-bound task concurrently with a single call.
• Three modes • async → native asyncio concurrency • thread → ThreadPoolExecutor under the hood • process → ProcessPoolExecutor for CPU-bound work
• Safe nesting – a single Runner can be reused recursively; the executor shuts down exactly once.
• Back-pressure – max_workers + semaphore keep concurrency under control.
• Sync and async entry points – call from “normal” code (run) or inside an event loop (arun).

Supported Python versions: 3.10 · 3.11 · 3.12 · 3.13

pip install conrun

import asyncio
from conrun import Runner


async def my_function(text: str) -> str:
    await asyncio.sleep(1)
    return text.upper()


async def main():
    result = await Runner(mode="async", max_workers=3).arun(
        my_function,
        ["a", "b", "c"],
    )
    print(result)  # ['A', 'B', 'C']


if __name__ == "__main__":
    asyncio.run(main())

MIT