open-webui/backend
Classic298 302b8cfd50
fix: tool calls make chats compact too early and inflate prompt token counts (#33218)
Since a recent refactor on dev, the prompt and completion token counts of a reply that used tools added up every model call in the loop, even though each later call already re-sends the whole conversation. A reply whose context was 50k tokens showed 90k, the context gauge read over 100% and the next message triggered compaction long before the chat reached the threshold. Prompt and completion tokens show the latest call again, while total tokens and cost still add up across all calls.
2026-10-26 19:17:34 +09:00
..
data refac: mv backend files to /open_webui dir 2024-09-04 16:54:48 +02:00
open_webui fix: tool calls make chats compact too early and inflate prompt token counts (#33218) 2026-10-26 19:17:34 +09:00
.dockerignore fix: litellm config issue 2024-02-24 22:35:11 -08:00
.gitignore refac 2024-09-06 04:59:20 +02:00
dev.sh perf: allow disabling websocket per-message-deflate (#28613) 2026-08-24 18:46:07 -04:00
requirements-slim.txt chore: bump Brotli to 1.2.1 and brotlicffi to 1.2.1.0 (#33073) 2026-10-25 04:49:42 +09:00
requirements.txt fix: pin preshed so the web loader keeps working (#33103) 2026-10-25 08:07:57 +09:00
start.sh chore: Python 3.12 and Debian 13 base image (#32861) 2026-10-22 04:51:57 +09:00
start_windows.bat chore: drop nltk, unused at the pinned versions (#29725) 2026-09-06 16:39:05 -04:00