truetype b15bd7794e fix(speech): generate TXT/JSON/MD locally from transcription result ...

Download buttons now use the already-completed transcription result
instead of re-sending the audio to the server for re-transcription.
This eliminates redundant server work and prevents download failures
when the second transcription attempt errors out.

Also wires up suppressNonSpeech and noSpeechThreshold options
through the API and server pipeline.

2026-09-16 21:22:41 +02:00

..

__pycache__

fix: resolve OpenVINO NPU and GPU backend switching/loading transition issues

2026-06-13 15:45:16 +05:30

build

rebrand to Uncensored AI Studio and implement automated backend build scripts and multi-platform setup utilities

2026-06-25 20:20:48 +05:30

config

implement modular backend setup scripts, new frontend components, and expanded server services

2026-06-22 10:10:29 +05:30

reset

add cross-platform reset scripts, portable server manager, and frontend scaffolding

2026-06-26 07:00:17 +05:30

search

implement modular web search capabilities with provider support, content fetching, and file-based caching.

2026-06-22 10:34:02 +05:30

server

fix(speech): generate TXT/JSON/MD locally from transcription result

2026-09-16 21:22:41 +02:00

setup

implement cross-platform tar extraction utility with symlink resolution fallback for setup scripts

2026-06-26 10:09:38 +05:30

workers

feat: optimize Apple NPU backend with safety checker detection, dynamic CFG enforcement, and improved model loading feedback

2026-06-25 06:53:34 +05:30