From 484fb6174381ccf0d8d7d11780166027fe7576df Mon Sep 17 00:00:00 2001
From: Classic298 <27028174+Classic298@users.noreply.github.com>
Date: Fri, 24 Jul 2026 01:17:41 +0200
Subject: [PATCH] perf: make streamed content accumulation genuinely linear
 (#27359)

The streaming handler's content accumulator is a closure cell (declared nonlocal in stream_body_handler), and CPython's in-place string append optimization only applies to plain local variables (STORE_FAST), never to cell variables (STORE_DEREF). The content += value form introduced in #27231 therefore still allocates and copies the full accumulated string on every delta, exactly like the f-string it replaced; whether that copy is cheap or expensive is up to the allocator, and measurements swing accordingly (6 to 83 ms of pure copying for a 400 KB response on Python 3.12, against 0.3 ms for this patch).

Accumulate the deltas in a list instead and join once at the single read site (publish_chat_finished_event at stream end). List append is amortized O(1) with no dependence on reference counts, bytecode specialization or allocator behaviour, so accumulation is O(n) by construction. The non-str fallback keeps the previous f-string coercion semantics.

Verified end to end against a mock SSE upstream: a streamed chat with a think-tag block plus 40 content deltas produces output items, message text, reasoning text and usage identical to current dev, with no errors in the server log.
---
 backend/open_webui/utils/middleware.py | 19 ++++++++-----------
 1 file changed, 8 insertions(+), 11 deletions(-)

diff --git a/backend/open_webui/utils/middleware.py b/backend/open_webui/utils/middleware.py
index 319407e4cf..5544888813 100644
--- a/backend/open_webui/utils/middleware.py
+++ b/backend/open_webui/utils/middleware.py
@@ -3974,9 +3974,10 @@ async def streaming_chat_response_handler(response, ctx):
             except Exception as e:
                 pass
 
-            content = (
+            initial_content = (
                 message.get('content', '') if message else last_assistant_message if last_assistant_message else ''
             )
+            content_parts = [initial_content] if initial_content else []
 
             # Initialize output: use existing from message if continuing, else create new
             existing_output = message.get('output') if message else None
@@ -3984,14 +3985,14 @@ async def streaming_chat_response_handler(response, ctx):
                 output = existing_output
             else:
                 # Only create an initial message item if there is content to initialize with
-                if content:
+                if initial_content:
                     output = [
                         {
                             'type': 'message',
                             'id': output_id('msg'),
                             'status': 'in_progress',
                             'role': 'assistant',
-                            'content': [{'type': 'output_text', 'text': content}],
+                            'content': [{'type': 'output_text', 'text': initial_content}],
                         }
                     ]
                 else:
@@ -4053,7 +4054,7 @@ async def streaming_chat_response_handler(response, ctx):
                     )
 
                 async def stream_body_handler(response, form_data):
-                    nonlocal content
+                    nonlocal content_parts
                     nonlocal usage
                     nonlocal output
                     nonlocal prior_output
@@ -4504,12 +4505,8 @@ async def streaming_chat_response_handler(response, ctx):
                                                 user,
                                             )
 
-                                        if isinstance(value, str):
-                                            # In-place append — avoids copying the full
-                                            # accumulated response on every chunk.
-                                            content += value
-                                        else:
-                                            content = f'{content}{value}'
+                                        # closure-cell str += recopies per chunk; append + join once at read is O(n)
+                                        content_parts.append(value if isinstance(value, str) else f'{value}')
 
                                         # Check if we're inside a tag-based block
                                         # (reasoning, code_interpreter, or solution).
@@ -5374,7 +5371,7 @@ async def streaming_chat_response_handler(response, ctx):
                             {'done': True},
                         )
 
-                await publish_chat_finished_event(request, user, metadata, title, content, output)
+                await publish_chat_finished_event(request, user, metadata, title, ''.join(content_parts), output)
 
                 await event_emitter(
                     {