aboutsummaryrefslogtreecommitdiff
path: root/app/backends
AgeCommit message (Collapse)Author
4 hoursfix: smart chunking, extraction, and process-safety issuesHEADmainhistoria
11 hoursfix: model-specific instructions support and set cfg-scale 4 for breeze ↵historia
automatically
5 daysfix: sglomni retry contract, booting detection, CLI keys, companion refresh, ↵historia
ffmpeg dep
5 daysrefactor: share the HF cache helpers between qwen and sglomnihistoria
6 daysfeat: pip install checks for bad crc errors and purges cache automaticallyhistoria
7 daysfix: increase gpu memory for MOSS with gplang-omni to .87historia
7 daysfeat: gplang-omni reserves 70% of gpu 0 memory, preventing oom errors on ↵historia
consumer cards
7 daysfix: warn before using a likely too-big chunk size for sglang-omni modelshistoria
7 daysfeat: reserve 20% vram when running higgs with sglang-omnihistoria
7 daysfeat: sglang-omni model picker on manual server launchhistoria
7 daysfix: ZONOS2 only generating 1024 tokens maxhistoria
7 daysfeat: auto-update via settingshistoria
8 daysfeat: sglang backend supporthistoria
9 daysslop refactor/deduphistoria
9 daysfix: spec santizer for glm, outetts, miotts, minimax.historia
9 daysfix: hide speech-to-speech only models from config wizardhistoria
10 daysfix: non-clone models correctly supported in tui, restart server when neededhistoria
12 daysfix: crash on bad model_specs from audio.cpp, sanitizedhistoria
12 daysfeat(audio.cpp): if github binary download fails, automatically start ↵historia
building from source
12 daysfeat: offer to download prebuilt mac/win binaries, new macos build fallback ↵historia
if xcode is not installed
12 daysfeat: cli auto-starts and stops locally-managed servers if no --api-url is ↵historia
passed
12 daysfeat: run auto-build audio.cpp scripts for macos and windowshistoria
13 daysrefactor: overhaul config.py, remove cli default optionshistoria
13 daysfix: remove backtrace from console output when using incorrect/incomplete ↵historia
cli flags
13 daysfeat: .wav hints for all voice directory pickers in tuihistoria
13 daysfix: system python used instead of venv python, breaking whisper transcriptionhistoria
13 daysremove: unnecessary flash on succesful buildhistoria
13 daysfeat: unified logging with logging_kit.pyhistoria
13 daysfeat: cuda arch detection for shorter builds, build failure detectionhistoria
13 daysfeat: qwen-tts backend widgets updated to same style as audio.cpphistoria
13 daysfix: silently failing whisper transcriptionhistoria
13 daysremove: unnecessary tui texthistoria
13 daysfeat: help menu in tuihistoria
13 daysfeat: automatic updates added to configure backend menuhistoria
2026-08-27feat: separate venvs for qwen-tts and faster, manage (un)installshistoria
2026-08-26feat: manage qwen-tts model installs manually, delete model(s) when uninstalledhistoria
2026-08-26feat: design model support for qwen-tts backend. remove unnecessary port ↵historia
split for qwen models
2026-08-26fix: settings menu only prompts to save after changehistoria
2026-08-26feat: language and option fields in tui, context-sensitive voice labelhistoria
2026-08-26feat: in-place toggle field, updated transcription tui to use new fieldhistoria
2026-08-26fix: default directory when choosing voices in tuihistoria
2026-08-26feat: combined install/configure tui screens into one menu, removed ↵historia
extraneous wizard screens
2026-08-26refactor: put shared ui screen code into ui.viewkithistoria
2026-08-26refactor: audiocpp.py setup flowhistoria
2026-08-26refactor: split tts.py into per-backend packageshistoria
2026-08-26fix: broken venv importshistoria
2026-08-26refactor(audiocpp): pin checkout to app/audio.cpp, report install resultshistoria
2026-08-26fix: harden server start/stop guardshistoria
2026-08-25feat: stop server and exit after generatinghistoria
2026-08-25fix: tui errors wait for getch()historia
generated by cgit v1.2.3 (git 2.25.1) at 2026-09-10 08:18:37 +0000