TIL: Concurrent API Calls With asyncio in Python

Processing audio files one at a time is slow. Each file means a network round trip to the API. asyncio runs multiple requests concurrently.

import asyncio
import aiohttp

async def transcribe(session, filepath):
    data = open(filepath, 'rb').read()
    async with session.post('https://api.example.com/v1/listen', data=data) as resp:
        return await resp.json()

async def main():
    files = ['file1.wav', 'file2.wav', 'file3.wav']
    async with aiohttp.ClientSession() as session:
        results = await asyncio.gather(
            *(transcribe(session, f) for f in files)
        )
    print(results)

asyncio.run(main())

asyncio.gather runs all the coroutines concurrently. The total time is the slowest single request, not the sum of all requests.

Do I need aiohttp or can I use requests?

requests is synchronous and blocks. Use aiohttp or httpx for async HTTP in Python.

What if one request fails?

gather raises the first exception by default. Use return_exceptions=True to collect results and exceptions separately.