Meet areq: Async Python Requests Without Losing Your Sanity

Ganesh Palanikumar — Tue, 08 Jul 2025 19:12:00 +0000

Like many of you, I’ve written my fair share of Python scripts that use requests. It’s simple. It works. And it has become almost muscle memory at this point.

The story starts with a simple piece of code that looks something like this.

import requests

def foo():
    try:
        response = requests.get("https://jsonplaceholder.typicode.com/todos/1")
        assert isinstance(response, requests.Response)
        response.raise_for_status()
        if response.ok:
            response_json = response.json()
            return response_json
    except requests.exceptions.RequestException as e:
        response_text = response.text
                print("Text:", response_text)
                return None

if __name__ == '__main__':
    foo()

It’s clean, concise, easily testable.

But then your application grows. Maybe you’re hitting hundreds of APIs, or building a web scraper . But now, your trusty synchronous code is no longer snappy. It starts to slow down. And someone suggests why not go async? The idea sticks in your head. It seems elegant, almost inevitable.

So, you start the refactor. You read up on httpx, the de-facto library for anyone wanting to make async requests. And it does seem simple - at first. But after refactoring and debugging, you are left with this.

import httpx
import asyncio

async def foo():
    try:
        async with httpx.AsyncClient() as client:
            response = await client.get("https://jsonplaceholder.typicode.com/todos/1")
            assert isinstance(response, httpx.Response)
            response.raise_for_status()
            if response.is_success:
                response_json = response.json()
                print(response_json)
    except httpx.HTTPError as e:
        print("Text:", getattr(response, 'text', None))
        return None

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

Great. You have joined the async revolution. But something feels weird. So many changes just to get the async flow to work!

httpx.AsyncClient() context manager
client.get() returns a coroutine that must be awaited
exceptions raised are no longer requests.exceptions.RequestException

From a refactoring perspective, I am definitely more nervous and less confident about the httpx code than the requests code. Not because httpx is a bad library, but because the affected area is a lot. If something goes wrong, it is going to be very hard to debug. Plus, I no longer recognize the code or the idioms.

This was exactly the problem I was trying to solve. A clean way to move from synchronous workflows using requests to asynchronous workflows using httpx without losing your sanity.

This led me to build areq: a minimal async drop-in replacement for requests. No weird new interfaces, no extra ceremony. Async where you want it. Familiarity where you expect it. The same code using areq would be like this.

import areq
import asyncio
import requests

async def foo():
    try:
        response = await areq.get("https://jsonplaceholder.typicode.com/todos/1")
        assert isinstance(response, requests.Response)
        response.raise_for_status()
        if response.ok:
            response_json = response.json()
            print(response_json)
    except requests.exceptions.RequestException as e:
        print("Text:", getattr(response, 'text', None))
        return None

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

That’s it. No extra configs, no new types to learn. No special idioms. Yes, I had to add async-await wherever required. But almost nothing else has changed. What’s more, isinstance(response, requests.Response) returns True!

No boilerplate context managers
No new Response object to learn
Type checks and except clauses remain same

Who did I build for?

The goal behind areq is simple: Bring async benefits to your software without wrecking your code (or your brain). You get the familiar ergonomics of requests, and the performance benefits of async. No rewrites, no API mental overhead.

areq is nothing fancy: it is just a simple wrapper over httpx. But I feel it will be really useful for some people.

🔄 Teams migrating to async gradually

You don’t always have the luxury to stop everything and refactor your entire codebase for async. With areq, you can start adopting async piece by piece, keeping the rest of your logic intact.

Want to convert just one function? One endpoint? One module? No problem.
areq enables you to do just that, without breaking .ok, type assertions or your test mocks.

🧠 People who eventually want to move to `httpx` (but not today)

areq isn’t here to compete with httpx.
In fact, it can be your on-ramp to httpx.

You can use areq to unlock async performance now—without rewriting everything. Later, once you’ve stabilized your logic and gathered profiling data, you can move to native httpx for deeper control or streaming needs.

🧵 Builders of CLI tools, background jobs, and lightweight APIs

If you’re working on scrapers, bots, micro services or any software that needs concurrency but don’t want the full mental load of switching to a new request paradigm, areq is for you. It gives you async without ceremony.

💼 Teams maintaining legacy requests code

Your old requests code works fine. It’s battle-tested.
Now you just want it to go faster.
areq lets you do that.
No rewrite. No breakage. No angry regression tests.

What’s Under the Hood?

areq:

Uses httpx.AsyncClient under the hood.
Converts the response to a requests.Response lookalike.
Converts common httpx exceptions (e.g. httpx.ConnectTimeout) into requests.exceptions.RequestException subclasses.
Supports all basic HTTP verbs

It’s not a full replacement for everything httpx or requests can do, but it covers 90% of practical use cases, especially when you’re trying to modernize an old codebase.

Known Limitations

Streaming and advanced session handling are not supported (yet!)
The response object is a subclass of requests.Response, but cannot be treated as same in all scenarios. But hey—it’s early. And it works for most real-world use cases.

Installing and Using `areq`

It is very easy to install areq. You can install it from PyPi just like any other python package.

pip install areq

How You Can Help

Star the repo

Starring the repo helps more people discover areq, and lets you keep up to date with features as they roll out.

Use it in your code!

Take areq for a spin. Use it in your code, tell me what breaks. File an issue.

Suggest features

Want streaming support? Cookie handling? What is that one feature that will make your life so much better?

Contribute!

areq is open sourced under MIT license. It is also my project with me as the sole developer. As such, if you have a cool idea, do fork the repository, implement and put up a PR. I will definitely review it.

Spread the word

Tell your friends and colleagues. Blog about it - the good, the bad and the ugly. The more people who use areq, the better it is going to become.

Final Thoughts

Sometimes all you need is a small wrapper that respects your old habits.
That’s areq. It won’t save the world, but it’ll save you a few afternoons.

If you’ve ever:

wanted to “just make it async” without rewriting everything
been burned by subtle differences between requests and httpx
Or you just want to make your trusty legacy code run a bit faster

Then areq might just be your new favorite micro-library.

DEV Community: Ganesh Palanikumar