Add decode_text parameter to WebSocket for receiving TEXT as bytes

[bdraco]

What do these changes do?

Adds a decode_text parameter to ws_connect() and WebSocketResponse() that allows receiving WebSocket TEXT messages as raw bytes instead of decoded strings.

When decode_text=False:

• TEXT messages return bytes instead of str
• Skips UTF-8 decoding overhead
• Enables direct use with high-performance JSON parsers like orjson that accept bytes

This addresses the use case where applications want to avoid the wasteful decode step when using JSON parsers that can process bytes directly.

Are there changes in behavior for the user?

No breaking changes. The default decode_text=True preserves existing behavior.

New behavior when decode_text=False:

• receive() returns WSMessageTextBytes (with bytes data) for TEXT messages
• receive_str() returns bytes instead of str
• receive_json() passes bytes to the loads function

Is it a substantial burden for the maintainers to support this?

Low burden. The implementation:

• Adds a single boolean flag passed through to WebSocketReader
• Uses existing message infrastructure with a new WSMessageTextBytes type
• Requires typing_extensions dependency for Python < 3.13

Why typing_extensions?

We use TypeVar with default= parameter (PEP 696, Python 3.13+) to make WebSocketResponse and ClientWebSocketResponse generic without breaking existing code. Without the default, all existing code would need to change from:

ws = WebSocketResponse()  # Would error without type parameter

to:

ws: WebSocketResponse[Literal[True]] = WebSocketResponse()

With default=Literal[True], the existing code continues to work unchanged while allowing explicit typing for decode_text=False:

ws: WebSocketResponse[Literal[False]] = WebSocketResponse(decode_text=False)

Related issue number

Fixes #11763

Checklist

• I think the code is well written
• Unit tests for the changes exist
• Documentation reflects the changes
• If you provide code modification, please add yourself to CONTRIBUTORS.txt
• Add a new news fragment into the CHANGES/ folder

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.75%. Comparing base (5d67dcf) to head (637d829).
⚠️ Report is 1 commits behind head on master.
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@           Coverage Diff            @@
##           master   #11764    +/-   ##
========================================
  Coverage   98.74%   98.75%            
========================================
  Files         127      127            
  Lines       43879    44171   +292     
  Branches     2337     2342     +5     
========================================
+ Hits        43327    43619   +292     
  Misses        392      392            
  Partials      160      160            

Flag	Coverage Δ
CI-GHA	98.60% <99.41%> (+<0.01%)	⬆️
OS-Linux	98.35% <99.41%> (+0.01%)	⬆️
OS-Windows	96.69% <97.05%> (+0.01%)	⬆️
OS-macOS	97.56% <97.05%> (+<0.01%)	⬆️
Py-3.10.11	97.12% <95.87%> (+<0.01%)	⬆️
Py-3.10.19	97.61% <98.23%> (+<0.01%)	⬆️
Py-3.11.14	97.80% <98.23%> (+<0.01%)	⬆️
Py-3.11.9	97.31% <95.87%> (-0.01%)	⬇️
Py-3.12.10	97.41% <95.87%> (-0.01%)	⬇️
Py-3.12.12	97.91% <98.23%> (+<0.01%)	⬆️
Py-3.13.11	98.15% <98.23%> (-0.01%)	⬇️
Py-3.14.2	98.17% <98.11%> (-0.01%)	⬇️
Py-3.14.2t	97.25% <95.61%> (-0.02%)	⬇️
Py-pypy3.11.13-7.3.20	97.42% <98.82%> (+0.01%)	⬆️
VM-macos	97.56% <97.05%> (+<0.01%)	⬆️
VM-ubuntu	98.35% <99.41%> (+0.01%)	⬆️
VM-windows	96.69% <97.05%> (+0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[bdraco]

now we have a problem with json.loads not decoding bytes

[bdraco]

Ok hundreds of lines of typing is not worth it for this change

[codspeed-hq]

CodSpeed Performance Report

Merging #11764 will not alter performance

_{Comparing no_decode_websocket_option (637d829) with master (5d67dcf)}

Summary

✅ 59 untouched

[Dreamsorcerer]

Should probably skip the 3.13 backport, as this is a significant new feature and will also likely have plenty of conflicts with new syntax etc.

[bdraco]

Thanks

[bdraco]

The backport is proving tricky. I'll merge this one I have more of it worked out locally

[patchback]

Backport to 3.14: 💔 cherry-picking failed — conflicts found

❌ Failed to cleanly apply 8b919d3 on top of patchback/backports/3.14/8b919d30a2339d3b9dbc53e8958f745b0f9703ea/pr-11764

Backporting merged PR #11764 into master

1. Ensure you have a local repo clone of your fork. Unless you cloned it
   from the upstream, this would be your origin remote.
2. Make sure you have an upstream repo added as a remote too. In these
   instructions you'll refer to it by the name upstream. If you don't
   have it, here's how you can add it:

   $ git remote add upstream https://github.com/aio-libs/aiohttp.git

3. Ensure you have the latest copy of upstream and prepare a branch
   that will hold the backported code:

   $ git fetch upstream
   $ git checkout -b patchback/backports/3.14/8b919d30a2339d3b9dbc53e8958f745b0f9703ea/pr-11764 upstream/3.14

4. Now, cherry-pick PR Add decode_text parameter to WebSocket for receiving TEXT as bytes #11764 contents into that branch:

   $ git cherry-pick -x 8b919d30a2339d3b9dbc53e8958f745b0f9703ea

   If it'll yell at you with something like fatal: Commit 8b919d30a2339d3b9dbc53e8958f745b0f9703ea is a merge but no -m option was given., add -m 1 as follows instead:

   $ git cherry-pick -m1 -x 8b919d30a2339d3b9dbc53e8958f745b0f9703ea

5. At this point, you'll probably encounter some merge conflicts. You must
   resolve them in to preserve the patch from PR Add decode_text parameter to WebSocket for receiving TEXT as bytes #11764 as close to the
   original as possible.
6. Push this branch to your fork on GitHub:

   $ git push origin patchback/backports/3.14/8b919d30a2339d3b9dbc53e8958f745b0f9703ea/pr-11764

7. Create a PR, ensure that the CI is green. If it's not — update it so that
   the tests and any other checks pass. This is it!
   Now relax and wait for the maintainers to process your pull request
   when they have some cycles to do reviews. Don't worry — they'll tell you if
   any improvements are necessary when the time comes!

🤖 @patchback
I'm built with octomachinery and
my source is open — https://github.com/sanitizers/patchback-github-app.