httpx follow redirects: why you get a 302 and how to fix it (tested)

httpx returns the 301/302 response instead of following it. Pass follow_redirects=True to the request or to httpx.Client(...) / AsyncClient(...); the redirect hops are then in r.history.

Unlike requests, httpx does not follow redirects by default: you get the 301/302 response itself. Measured with httpx 0.28.1 (and requests 2.34.2 for comparison) on Python 3.13.5 against a local http.server (srv.start() starts it in a thread; /r302 redirects to /final, which echoes the method).

1. httpx returns the redirect instead of following it

import httpx, requests
from srv import start
base = start()

r = requests.get(base + "/r302")
print("requests:", r.status_code, r.url)

r = httpx.get(base + "/r302")
print("httpx:   ", r.status_code, r.url)
print("is_redirect:", r.is_redirect)
print("location:", r.headers["location"])
print("next_request:", r.next_request)

Output:

requests: 200 http://127.0.0.1:8963/final
httpx:    302 http://127.0.0.1:8963/r302
is_redirect: True
location: /final
next_request: <Request('GET', 'http://127.0.0.1:8963/final')>

requests ends on /final with 200; httpx stops at the 302. The response tells you where it would go: r.is_redirect, the Location header, and r.next_request (a ready-made request you can send with client.send(r.next_request)).

2. Fix: follow_redirects=True (per request or on the Client)

import httpx
from srv import start
base = start()

# per request
r = httpx.get(base + "/r302", follow_redirects=True)
print(r.status_code, r.url)
print("history:", r.history)

# or once, on the client
with httpx.Client(follow_redirects=True) as client:
    r = client.get(base + "/r301")
    print(r.status_code, r.json()["path"], [h.status_code for h in r.history])

Output:

200 http://127.0.0.1:8963/final
history: [<Response [302 Found]>]
200 /final [301]

The redirect responses are kept in r.history, and r.url is the final URL. The old allow_redirects= argument (requests, and httpx before 0.20) is not accepted by httpx 0.28; use follow_redirects.

3. raise_for_status() raises on a redirect

import httpx
from srv import start
base = start()

r = httpx.get(base + "/r302")
try:
    r.raise_for_status()
except httpx.HTTPStatusError as e:
    print(type(e).__name__)
    print(e)

Output:

HTTPStatusError
Redirect response '302 Found' for url 'http://127.0.0.1:8963/r302'
Redirect location: '/final'
For more information check: https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/302

If you call raise_for_status() without following redirects, a 3xx is an error too (only 2xx passes). The message names the redirect location, which is usually the clue that follow_redirects is missing.

4. Redirect loops and POST becoming GET

import httpx
from srv import start
base = start()

with httpx.Client(follow_redirects=True, max_redirects=5) as client:
    try:
        client.get(base + "/loop")
    except httpx.TooManyRedirects as e:
        print(type(e).__name__, "-", e)

    for code in (301, 302, 303, 307, 308):
        r = client.post(base + f"/r{code}", json={"a": 1})
        print(code, "->", r.json()["method"])

Output:

TooManyRedirects - Exceeded maximum allowed redirects.
301 -> GET
302 -> GET
303 -> GET
307 -> POST
308 -> POST

A loop raises httpx.TooManyRedirects after max_redirects (default 20). When following, a POST answered with 301/302/303 is re-sent as GET (body dropped), while 307 and 308 keep the POST. If your POST "loses" its data after a redirect, the server is sending 301/302 (often a missing trailing slash or http→https); post to the final URL instead.

5. AsyncClient works the same way

import asyncio, httpx
from srv import start
base = start()

async def main():
    async with httpx.AsyncClient(follow_redirects=True) as client:
        r = await client.get(base + "/r302")
        print(r.status_code, r.url)

asyncio.run(main())

Output:

200 http://127.0.0.1:8963/final

Related: 301 vs 302 vs 307 vs 308, check a redirect chain, ERR_TOO_MANY_REDIRECTS, requests HTTPError response message.

Written with AI assistance. Measured on 2026-10-23 with httpx 0.28.1 and requests 2.34.2 on Python 3.13.5 against a local http.server. Each block was run verbatim and the output pasted unedited. PlainHeaders is our extension; httpx is free.