plimit

Run multiple async functions with limited concurrency

Python port of the popular Node.js p-limit package. Uses asyncio to run async (and sync) functions with a configurable concurrency limit.

Install

pip install plimit

or with uv:

uv add plimit

Usage

import asyncio
from plimit import p_limit

async def main():
    limit = p_limit(2)  # max 2 concurrent tasks

    async def fetch(url):
        # ... your async work
        return url

    results = await asyncio.gather(
        limit(fetch, "https://example.com/1"),
        limit(fetch, "https://example.com/2"),
        limit(fetch, "https://example.com/3"),
    )
    print(results)

asyncio.run(main())

API

p_limit(concurrency, *, reject_on_clear=False)

Returns a Limiter instance.

• concurrency (int) — Maximum number of async functions running at the same time. Minimum: 1.
• reject_on_clear (bool) — When True, clear_queue() rejects pending futures with ClearQueueError instead of silently discarding them. Default: False.

limit(fn, *args) → awaitable

Schedule fn(*args) to run under the concurrency limit. Returns an awaitable that resolves to the return value of fn.

• fn can be an async function or a regular function.
• Extra positional arguments are forwarded to fn.

result = await limit(my_async_fn, arg1, arg2)

limit.active_count → int

Number of functions currently executing.

limit.pending_count → int

Number of functions waiting in the queue.

limit.concurrency → int (get/set)

Get or set the concurrency limit at runtime. When increased, queued tasks start immediately up to the new limit.

limit.concurrency = 10  # increase concurrency dynamically

limit.clear_queue()

Discard pending functions that have not started yet.

• If reject_on_clear=True, pending futures are rejected with ClearQueueError.
• Does not cancel functions that are already running.

limit.map(iterable, fn) → awaitable list

Process an iterable through fn with limited concurrency. The mapper receives (item, index).

results = await limit.map([1, 2, 3], async_transform)

limit_function(fn, concurrency, *, reject_on_clear=False)

Convenience wrapper — returns a new async function that calls fn with limited concurrency.

from plimit import limit_function

limited_fetch = limit_function(fetch, concurrency=3)
result = await limited_fetch(url)

ClearQueueError

Exception raised for pending tasks when clear_queue() is called with reject_on_clear=True.

Comparison with JavaScript p-limit

Requirements

Python 3.10+

License

MIT