-
e3b2a53c54
Self-restart worker process when model loading fails
main
du5t
2026-07-24 00:22:59 +09:00
-
89f8643426
Pin remaining unbounded deps, drop stale JS model-list fallback, add smoke test
du5t
2026-07-23 23:29:30 +09:00
-
9472387d1b
Auto-unload idle models and evict on model switch across all workers
du5t
2026-07-23 23:29:13 +09:00
-
ce064d6894
Add VibeVoice-ASR backend
du5t
2026-07-23 16:54:34 +09:00
-
a32b0acb09
Fix broken file transcription and speaker diarization
du5t
2026-07-23 16:31:55 +09:00
-
22e9b805a6
Persist realtime transcription sessions to history
du5t
2026-06-19 00:30:15 +09:00
-
acdf098c41
Add XTTS-v2 as first TTS backend, extensible for more
du5t
2026-06-18 18:46:45 +09:00
-
ac49bfc7f9
Cut over OIDC redirect to speech.du5t.xyz
du5t
2026-06-18 17:53:05 +09:00
-
665ff8a659
Rename project from asr-v2 to speech
du5t
2026-06-18 15:25:56 +09:00
-
6b9a9cce8f
Add Authentik OIDC login
du5t
2026-06-18 13:44:20 +09:00
-
cabbdac39b
Restructure into asr/tts packages and add TTS tab skeleton
du5t
2026-06-18 11:28:43 +09:00
-
ab6df14044
Fix volume SELinux labels and build.sh deploy instructions
du5t
2026-06-11 15:32:20 +09:00
-
a618f59820
Enable GPU passthrough in quadlet
du5t
2026-06-10 18:01:59 +09:00
-
56e637a300
Initial commit: ASR v2 with multi-venv isolation
du5t
2026-05-23 00:24:36 +09:00