velixio 090cc37144 fix(memory): release the model before emptying the cache, not after ...

The shared voice model's unload emptied the allocator caches and *then*
dropped the reference. That frees nothing: the weights are still reachable
when gc.collect() runs, empty_cache() only returns blocks the allocator
already considered free, and the reference drops a moment later into a cache
nothing will flush again. The unload logs success, the engine leaves the
registry, and nvidia-smi does not move.

Six modules open-coded the same two lines. Exactly one had them inverted --
OmniVoiceBackend.unload, which is the path the engine-registry idle sweep
reaches, which is the sweep a headless worker node runs. So every unload a
user could trigger from the UI worked, and the one that runs unattended on a
machine lending its GPU held 3.6 GB indefinitely. Found on hardware: the
sweep fired on schedule, logged "Released 1 idle engine(s)", and VRAM stayed
flat at 3656 MiB for the next two minutes.

Replace all six with model_manager.unload_shared_model(), which clears the
reference, drops the clone-prompt side cache, then frees -- in that order,
in one place. Two callers gain the side-cache drop they were missing
(/system/flush-memory and the shutdown path), which is the same defect one
step down: an unload that kept the encoded reference tensors belonging to the
model it had just released.

A source guard asserts nothing outside model_manager assigns the shared
reference, so the next caller cannot reintroduce the ordering. It caught the
sixth site while being written.

Also give the AudioSeal watermark models the bargain every other model in the
app already makes: they loaded on the first embed and stayed resident for the
life of the process. CPU-resident, so this is system RAM rather than VRAM,
and the machines that notice are the ones running batches.

The error text on a failing unload changes with the ordering. "Could not be
unloaded, retry after the current generation finishes" was accurate when the
cache flush ran first and aborted before the release; now the release has
already happened and only the flush can fail, so it says that instead of
sending the user to repeat work that is done.

2026-08-12 02:33:04 +05:30

..

backend

feat(indextts): add native IndexTTS 2.5 support (#1485)

2026-08-11 01:26:25 +00:00

evals

test(evals): LLM-judge eval tier — non-gating semantic suites (Wave 0.3) (#355)

2026-06-11 22:00:30 +05:30

fixtures

feat(workers): make inbound mode reachable — settings, endpoints, docs

2026-08-11 22:28:06 +05:30

frontend

Release VoiceStudio 5.0.0 (#1487)

2026-08-11 04:05:45 +00:00

probe

fix(dictation): stop the widget stranding an empty square, repair the swept data dirs (#1398)

2026-08-07 05:39:41 +05:30

scripts

fix(dictation): stop the widget stranding an empty square, repair the swept data dirs (#1398)

2026-08-07 05:39:41 +05:30

smoke

feat: TTS-only first run, platform-curated ASR, guided OS permissions, parakeet-mlx

2026-07-17 15:20:54 +05:30

conftest.py

test: stop config-path leakage across test modules (#1269) (#1307)

2026-07-29 12:23:52 -07:00

test_agentic_provider_contract.py

fix(api): /v1/audio/speech honors num_step + guidance_scale instead of silently dropping them (#1014) (#1035)

2026-07-09 17:49:25 +05:30

test_analytics_lifecycle.py

feat(analytics): in-repo publishable token — source builds get the same consent-gated analytics (#1193)

2026-07-20 11:51:40 +05:30

test_analytics_optin.py

feat(analytics): in-repo publishable token — source builds get the same consent-gated analytics (#1193)

2026-07-20 11:51:40 +05:30

test_api_route_inventory.py

test: backend route-inventory + webUI feature-coverage guards (#609)

2026-06-22 05:33:00 +05:30

test_api.py

fix(security): move host-path authority into the desktop shell (#1448)

2026-08-09 21:55:43 +00:00

test_app_version.py

chore(version): make frontend/package.json the single source of truth (#503)

2026-06-16 22:50:39 +05:30

test_apprun_launcher.py

fix(linux): AppRun workaround detection reads the BUNDLED WebKitGTK version, not the host's (#961 follow-up) (#1024)

2026-07-09 03:53:46 +05:30

test_archetype_blank_guard.py

fix(audio): stop near-silent renders becoming blank noise + guard archetype renders (#204)

2026-05-31 13:27:30 +05:30

test_asr_device_aware_autodetect.py

feat: TTS-only first run, platform-curated ASR, guided OS permissions, parakeet-mlx

2026-07-17 15:20:54 +05:30

test_asr_execstack_fallback_692.py

fix(asr): fall back instead of crashing when CTranslate2's .so won't load (#692) (#699)

2026-06-26 00:04:54 +05:30

test_asr_ffmpeg_decode.py

fix(asr): decode WhisperX audio via validated ffmpeg, not bare PATH lookup (#479) (#482)

2026-06-16 13:26:26 +05:30

test_asr_gpu_compat.py

feat: TTS-only first run, platform-curated ASR, guided OS permissions, parakeet-mlx

2026-07-17 15:20:54 +05:30

test_asr_model_missing.py

fix(security): stabilize inner transcription failures

2026-08-09 23:35:28 +00:00

test_asr_openai_compat_877.py

feat(asr): complete activation flow for the OpenAI-compatible remote ASR engine (#1087)

2026-07-12 02:40:45 +05:30

test_asr_vram_preflight.py

fix(asr): VRAM preflight before whisperx load — no more native OOM abort on 8 GB cards (#723) (#870)

2026-07-02 04:48:51 +05:30

test_assign_speakers_from_diarization.py

feat(dub): optional speaker-count hint for diarization (#274) (#275)

2026-06-03 14:10:08 +05:30

test_assign_speakers_from_turns.py

feat(dub): FunASR cam++ inline diarization → dub speakers (#182 Phase 2) (#197)

2026-05-31 05:43:33 +05:30

test_audiobook_cancel.py

feat(audiobook): cast/voice mapping + markup toolbar + live stats + validation (#1217)

2026-07-21 01:54:13 +05:30

test_audiobook_cast.py

feat(audiobook): cast/voice mapping + markup toolbar + live stats + validation (#1217)

2026-07-21 01:54:13 +05:30

test_audiobook_cover.py

feat(audiobook): metadata, cover art, format + loudness UI (PR 2/8) (#409)

2026-06-13 14:02:10 +05:30

test_audiobook_expressive.py

fix(audiobook): address #1208 review — bound expressive knobs, canonical markup tokens in all locales

2026-07-20 22:44:56 +05:30

test_audiobook_preview.py

fix(watermark): route every synthetic-audio producer through one mark_synthetic chokepoint (#1169)

2026-07-16 23:08:36 +05:30

test_audiobook_remote.py

feat(workers): remote downloads, audiobook chapters, and one port that stays honest

2026-08-11 16:53:33 +05:30

test_audiobook.py

fix(audiobook): chapter render no longer crashes on mixed 1-D/2-D audio chunks (#897) (#953)

2026-07-05 00:40:35 +05:30

test_autofit_838.py

fix(translate): Cinematic/Autofit can no longer invent dialogue — divergence guard + pinned temperature (#950)

2026-07-04 22:28:37 +05:30

test_backend_marker_header_1385.py

fix(client): a foreign 404 page is a routing problem, not a backend error (#1386)

2026-08-06 15:13:08 +05:30

test_bearer_middleware.py

feat(remote): backend URL + bearer key + Tailscale docs (Wave 2.3) (#364)

2026-06-12 03:57:20 +05:30

test_bind_host.py

P0: release.yml typecheck + bind audit + loopback middleware (#84)

2026-05-18 22:00:59 +05:30

test_broken_dependency_diagnosis_1415.py

fix(models): a broken dependency is not a missing install, and is not silent (#1415) (#1423)

2026-08-08 05:47:45 +05:30

test_capture_asr_idle_release.py

fix(memory): idle-release the dictation ASR — the real cause of the 16 GB OOM deaths (#1104)

2026-07-12 14:38:14 +05:30

test_capture_refine.py

feat: TTS-only first run, platform-curated ASR, guided OS permissions, parakeet-mlx

2026-07-17 15:20:54 +05:30

test_capture_ws.py

feat: TTS-only first run, platform-curated ASR, guided OS permissions, parakeet-mlx

2026-07-17 15:20:54 +05:30

test_changelog_parse.py

feat(updates): data-safe updates — pre-migration DB backups, guarded venv heal, release notes + changelog reader (#909)

2026-07-02 23:52:17 +05:30

test_changelog_style.py

feat(workers): remote GPU workers over a versioned gRPC protocol

2026-08-10 14:18:42 +05:30

test_chunk_progress_extends_generate_budget_1391.py

fix(generate): a render that keeps finishing chunks is not wedged (#1392)

2026-08-06 16:30:31 +05:30

test_chunked_tts_dense_script.py

fix(tts): NFC-normalize text + dense-script-aware chunking for long-form quality (#502/#505) (#587)

2026-06-22 03:18:15 +05:30

test_chunked_tts.py

fix(audiobook): chapter render no longer crashes on mixed 1-D/2-D audio chunks (#897) (#953)

2026-07-05 00:40:35 +05:30

test_ci_windows_ffmpeg_retry.py

ci(windows): make the ffmpeg retry test the outcome, not choco's exit code (#1290)

2026-07-29 01:24:27 -07:00

test_classify_401_substring.py

test(errors): pin audio failure class with 401 path

2026-08-09 20:23:05 +00:00

test_clone_prompt_cache.py

perf(dub): single-use per-segment refs no longer evict the prompts a dub reuses; add docs/performance.md (#1132)

2026-07-13 14:36:06 +05:30

test_clone_prompt_wiring.py

perf(dub): single-use per-segment refs no longer evict the prompts a dub reuses; add docs/performance.md (#1132)

2026-07-13 14:36:06 +05:30

test_clone_ref_silence_ladder.py

fix: resolve open issue batch — #1172 #1173 #1174 #1185 #1186 #1188

2026-07-19 04:29:33 +05:30

test_clone_refs_live_in_their_own_job_dir.py

fix(dub): clone references live in their own job dir; deletes spare shared files (#1331) (#1381)

2026-08-05 21:47:15 +05:30

test_compile_fallback.py

fix(tts): torch.compile failures fall back to eager — generation never fails on unsupported GPUs (#278) (#327)

2026-06-11 13:09:18 +05:30

test_compile_thread_affinity.py

fix(tts): pin cudagraph-compiled model inference to one dedicated thread (#315) (#343)

2026-06-11 13:58:21 +05:30

test_confucius4_scaffold.py

fix(engines): Confucius4-TTS validated E2E — clone sys.path import, 22.05 kHz, real install docs (#590) (#872)

2026-07-02 04:49:50 +05:30

test_confucius4_sidecar.py

fix(engines): Confucius4-TTS validated E2E — clone sys.path import, 22.05 kHz, real install docs (#590) (#872)

2026-07-02 04:49:50 +05:30

test_content_disposition_unicode.py

fix(errors): six failures that reached users as raw text (#1262, #1256, #1251, #1247, #1257, #1254) (#1264)

2026-07-26 12:39:29 -07:00

test_corrupt_weights_recovery_1406.py

Merge main and tighten tokenizer recovery regression

2026-08-10 04:55:58 +00:00

test_cuda_arch_compat.py

fix(cuda): stop sending every RTX 40-series card to the CPU (#1289)

2026-07-29 01:59:06 -07:00

test_cudnn8_hard_crash_1371.py

fix(asr): stop a missing cuDNN 8 from killing the backend process (#1371)

2026-08-07 05:24:05 +05:30

test_db_backup.py

feat(updates): data-safe updates — pre-migration DB backups, guarded venv heal, release notes + changelog reader (#909)

2026-07-02 23:52:17 +05:30

test_db_migration_safety.py

fix: restore generation and startup broken by the rename (#1417) (#1420)

2026-08-08 02:40:52 +05:30

test_db_schema_reconcile.py

fix(db): self-heal additive schema columns so consent_audio_path 500s stop (#557)

2026-06-20 09:00:47 +05:30

test_degraded_api_observability.py

Merge main and address degraded-state review findings

2026-08-10 09:21:54 +00:00

test_deleting_a_dub_keeps_dirs_others_use.py

fix(dub): clone references live in their own job dir; deletes spare shared files (#1331) (#1381)

2026-08-05 21:47:15 +05:30

test_dependency_security_floors.py

Merge remote-tracking branch 'origin/main' into fix/dependabot-security

2026-08-10 04:30:23 +00:00

test_describe_voice.py

feat(design): free-text 'describe your voice' field maps to design parameters (#317) (#331)

2026-06-11 12:14:45 +05:30

test_desktop_prod_scripts.py

test: gate AppImage execution contract to Linux

2026-08-11 13:22:10 +00:00

test_destructive_cleanup.py

test: isolate destructive cleanup imports

2026-08-10 07:33:49 +00:00

test_device_caps.py

feat(routing): host device probe + routing resolver (#21 PR 1/5) (#430)

2026-06-13 23:03:22 +05:30

test_device_unsupported_fallback.py

fix(device): fall back to CPU when the GPU arch is unsupported, not 500 every generate (#756) (#757)

2026-07-01 17:59:49 +05:30

test_diagnose.py

feat: rename the product to VoiceStudio (previously OmniVoice-Studio)

2026-08-07 01:30:58 +05:30

test_diagnostic_bundle.py

feat(setup): unified first-run journey — install gate, studio-console wizard, platform awareness (#295)

2026-06-11 12:15:28 +05:30

test_diarization_error_class.py

fix: speaker detection — gated pyannote license surfaces a docs deeplink (closes #78) (#110)

2026-05-20 14:08:55 +05:30

test_diarization_weights_only.py

fix(diarization): register torch safe-globals before pyannote load (#270) (#271)

2026-06-03 10:37:37 +05:30

test_dictation_no_pill_window.py

fix(dictation): stop the widget window stranding a rectangle, and stop showing a pill

2026-08-07 17:41:43 +05:30

test_dictation_router.py

Merge current main into GHAS empty-except fixes

2026-08-10 11:13:10 +00:00

test_dots_tts.py

feat(tts): add MOSS-TTS-v1.5 (8B) and dots.tts (2B) as opt-in engines (#498)

2026-06-20 00:02:27 +05:30

test_dropped_audio_reaches_the_user_1330.py

feat(workers): one gateway for GPU calls, and results too big for the wire

2026-08-11 13:30:26 +05:30

test_dropped_chunk_is_not_silent.py

fix(tts): a chunk that renders to nothing is no longer dropped in silence (#1330) (#1360)

2026-08-04 15:08:36 +05:30

test_dub_audio_export.py

fix(dub): hydrate partial translations on tab switch; dialect guard moves into the store (#1149)

2026-07-14 02:11:43 +05:30

test_dub_batch_engine_selection.py

fix(watermark): route every synthetic-audio producer through one mark_synthetic chokepoint (#1169)

2026-07-16 23:08:36 +05:30

test_dub_dialect.py

fix(translate): Cinematic/Autofit can no longer invent dialogue — divergence guard + pinned temperature (#950)

2026-07-04 22:28:37 +05:30

test_dub_download_403.py

fix(dub+generate): yt-dlp 403 player-client fallback (#625) + non-finite audio guard (#629) (#635)

2026-06-22 23:36:32 +05:30

test_dub_download_ffmpeg_location.py

fix(dub): pass OmniVoice's ffmpeg to yt-dlp so URL merge works off PATH (#712) (#716)

2026-06-26 15:16:45 +05:30

test_dub_download_mtime.py

fix(dub): skip yt-dlp mtime stamp to avoid [Errno 22] on Windows (#642) (#644)

2026-06-23 11:53:44 +05:30

test_dub_download_os_error.py

fix(dub): #1225 review — the preflight error had no class, and ENOENT no facts

2026-07-23 03:53:30 +05:30

test_dub_download_retry.py

fix(dub): retry transient broken-pipe on URL download (#579, #598) (#605)

2026-06-22 05:09:51 +05:30

test_dub_duration_plan_api.py

fix(watermark): route every synthetic-audio producer through one mark_synthetic chokepoint (#1169)

2026-07-16 23:08:36 +05:30

test_dub_error_transparency.py

fix(dub): stereo, full-band music bed — separate the HQ extraction, pin the mix to stereo (#1138)

2026-07-13 21:04:06 +05:30

test_dub_export_bitrate.py

test: add preflight + bitrate coverage, refresh legacy mocks, wire CI gate

2026-04-22 18:46:01 +05:30

test_dub_export_unique.py

fix: close remaining dub export path sinks

2026-08-10 08:57:27 +00:00

test_dub_job_deleted_mid_ingest.py

fix(dub): purge order was hash-dependent, and the cap could evict a live marker (#1271)

2026-07-27 12:22:18 -07:00

test_dub_multispeaker_voice_486.py

fix(watermark): route every synthetic-audio producer through one mark_synthetic chokepoint (#1169)

2026-07-16 23:08:36 +05:30

test_dub_no_tts_load_for_asr.py

fix(desktop): make recording and dubbing reliable (#1481)

2026-08-10 22:43:37 +00:00

test_dub_onsets_route.py

test: pay down export-router test debt + kill two test-order pollution classes (#1081)

2026-07-11 23:10:35 +05:30

test_dub_parse_subtitle_text.py

feat(dub): paste a translation from an external source onto existing segments

2026-07-20 13:18:51 +05:30

test_dub_per_lang_subtitles.py

feat(dub): per-language translations + per-track caches — switching languages stops destroying work (P1) (#958)

2026-07-05 03:05:40 +05:30

test_dub_pipeline_state.py

fix(dub): completed tracks always show their tabs + history keeps its language (P0) (#956)

2026-07-05 01:21:23 +05:30

test_dub_qc.py

feat(dub): second-pass ASR QC — flag lines whose dub drifts from target (Wave 3.3) (#370)

2026-06-12 12:50:39 +05:30

test_dub_remote_safety.py

feat(workers): remote downloads, audiobook chapters, and one port that stays honest

2026-08-11 16:53:33 +05:30

test_dub_remote.py

feat(workers): dubbing goes remote, and the protocol stops lying to old workers

2026-08-11 17:39:46 +05:30

test_dub_seg_path.py

fix(dub): key per-segment WAVs by stable id, not list index (#185) (#187)

2026-05-31 03:31:03 +05:30

test_dub_subtitles_309.py

fix(errors): six failures that reached users as raw text (#1262, #1256, #1251, #1247, #1257, #1254) (#1264)

2026-07-26 12:39:29 -07:00

test_dub_timing_strategy.py

feat: onboarding demos, opt-in bug reporting, error-docs deeplinks + issue triage (#133)

2026-05-29 17:24:25 +05:30

test_dub_transcribe.py

fix(desktop): make recording and dubbing reliable (#1481)

2026-08-10 22:43:37 +00:00

test_dub_translate.py

fix(translate): Dub LLM engine runs on the configured LLM provider (new dub_translation skill) (#944)

2026-07-04 19:22:12 +05:30

test_dub_translation_quality.py

feat(dub): LLM translations keep terms consistent and sound spoken — auto-glossary brief + reflect pass (#1050)

2026-07-10 18:26:07 +05:30

test_dub_voice_match.py

fix(dub): restore cast voices from source audio (#1484)

2026-08-10 23:46:37 +00:00

test_duration_nfc.py

fix(lang): propagate profile/request language into generation + longform (#533/#505/#502) (#565)

2026-06-20 14:45:58 +05:30

test_duration_planner.py

feat(dub): predict segment fit before synthesis — tight/impossible badges + opt-in shorter rewrites (#1051)

2026-07-10 18:40:16 +05:30

test_effects_chain.py

Per-segment audio effects DSP preset selector (closes #67, rebased from #68) (#109)

2026-05-20 13:40:16 +05:30

test_endpoint_race.py

fix(security): constrain endpoint probes to trusted origins (#1447)

2026-08-09 21:18:11 +00:00

test_engine_language_rejection.py

fix(errors): six failures that reached users as raw text (#1262, #1256, #1251, #1247, #1257, #1254) (#1264)

2026-07-26 12:39:29 -07:00

test_engine_memory.py

feat(memory): one TTS engine resident at a time — stop stacking models on 16 GB (#1105)

2026-07-12 15:05:11 +05:30

test_engine_routing.py

fix(engines): revive dead license dialog, refresh matrix on select, surface routing verdict (#905)

2026-07-02 23:22:49 +05:30

test_engine_venv_probe_1414.py

fix(engines): a slow venv probe is not a broken venv (#1414) (#1421)

2026-08-08 05:27:47 +05:30

test_engines.py

fix(mlx): pass the voice description the design model requires (#1405)

2026-08-07 16:15:31 +05:30

test_error_journal.py

fix: keep degraded state truthful

2026-08-10 08:22:37 +00:00

test_exports_api.py

fix(security): authorize export filesystem sinks

2026-08-10 07:32:50 +00:00

test_failure_ansi_stripping.py

fix(errors): strip terminal colour codes before a failure reaches the user (#1344) (#1353)

2026-08-04 11:26:11 +05:30

test_failure_classify.py

fix(errors): the transformers-import advice names torchvision now (#1376) (#1377)

2026-08-05 17:39:19 +05:30

test_failure_helper.py

feat: pipeline error transparency — no more silent "unknown error" (plan-04, closes #131) (#136)

2026-05-29 09:12:23 +05:30

test_fdl_download_preflight.py

Fast model downloads: Xet fast path + accurate progress (FDL W0–W2 + W4) (#424)

2026-06-13 20:15:15 +05:30

test_fdl_segmented_download.py

Fast model downloads: Xet fast path + accurate progress (FDL W0–W2 + W4) (#424)

2026-06-13 20:15:15 +05:30

test_ffmpeg_banner_stripping.py

fix(errors): strip ffmpeg's banner so the failure is the message (#1311)

2026-07-29 15:02:02 -07:00

test_ffmpeg_failure_diagnosis.py

fix(dub): diagnose export failures honestly; dodge the Windows 32K argv limit (#1152)

2026-07-16 01:45:23 +05:30

test_filesystem_boundaries.py

fix: close remaining dub export path sinks

2026-08-10 08:57:27 +00:00

test_fit_planner.py

feat(dub): underrun fill — short dubbed lines are slowed toward their slot instead of leaving dead air (#1137)

2026-07-13 19:09:02 +05:30

test_fitted_subtitles.py

feat(dub): regenerate subtitle timeline on the fitted timeline (Wave 3.1) (#371)

2026-06-12 13:03:00 +05:30

test_funasr_backend.py

fix(asr): restore FunASR SenseVoice diarization (#1167)

2026-07-16 22:59:34 +05:30

test_funasr_global_speaker_chunking.py

feat: TTS-only first run, platform-curated ASR, guided OS permissions, parakeet-mlx

2026-07-17 15:20:54 +05:30

test_gallery_previews.py

feat(workers): remote downloads, audiobook chapters, and one port that stays honest

2026-08-11 16:53:33 +05:30

test_generate_engine.py

feat(routing): synth-time no-silent-fallback gating at all TTS entry points (#21 follow-up) (#440)

2026-06-14 11:18:37 +05:30

test_generate_ref_text_persist.py

fix(studio): restore Clear History, stoppable auto-play preview, + cached ref transcripts (#1032) (#1036)

2026-07-09 18:43:33 +05:30

test_generate_streaming.py

fix(security): keep stream failures out of logs

2026-08-09 23:18:16 +00:00

test_generate_timeout_730.py

fix(diagnostics): TTS generate timeout message tells you to Flush/Unload (#939) (#999)

2026-07-08 05:18:00 +05:30

test_generation_audio_guard.py

fix(errors): stop giving advice that cannot work (#1347, #1335, #1334) (#1374)

2026-08-05 14:45:29 +05:30

test_generation_history_self_heal.py

fix(generate): self-heal schema + don't 500 a generated clip on a history-write fail (#710) (#714)

2026-06-26 15:04:31 +05:30

test_generation_remote_routing.py

feat(workers): dubbing goes remote, and the protocol stops lying to old workers

2026-08-11 17:39:46 +05:30

test_generation_takes.py

feat(studio): generation takes — star, replay, and restore past takes; capped history retention (#1052)

2026-07-10 18:54:14 +05:30

test_generation_timeout_is_named.py

fix(generate): a deadline is not an unrecognized error (#1368) (#1373)

2026-08-05 02:43:29 +05:30

test_gguf_source_build_runtime_1348.py

fix(gguf): a source-built runtime must actually run (#1348) (#1384)

2026-08-06 04:32:40 +05:30

test_ghas_false_positive_invariants.py

test(security): lock reviewed GHAS invariants

2026-08-09 21:48:54 +00:00

test_gitleaks_allowlist.py

test(security): derive public analytics allowlist

2026-08-10 07:12:56 +00:00

test_glossary_scrub.py

feat(settings): LLM Skills — per-feature enable/route control for every LLM call (#912)

2026-07-03 00:08:30 +05:30

test_gptsovits_endpoint_security.py

fix(security): constrain GPT-SoVITS endpoints (#1463)

2026-08-10 11:29:42 +00:00

test_gpu_gateway_scheduler_contract.py

feat(workers): remote downloads, audiobook chapters, and one port that stays honest

2026-08-11 16:53:33 +05:30

test_gpu_gateway.py

feat(workers): dubbing goes remote, and the protocol stops lying to old workers

2026-08-11 17:39:46 +05:30

test_gpu_pool_queue_accounting_1190.py

fix(gpu-pool): sanitize the job label before logging (CodeQL)

2026-07-20 14:35:21 +05:30

test_gpu_pool_resilient.py

fix(backend): self-healing GPU pool so a reset can't strand requests (#589/#599) (#601)

2026-06-22 04:42:16 +05:30

test_gpu_routing_verdict.py

feat(routing): active-engine GPU verdict in preflight + diagnose (#21 PR 4/5) (#433)

2026-06-14 02:07:54 +05:30

test_gpu_wedge_stack_capture.py

fix(gpu): record WHERE a wedged job was stuck when its budget expires (#1338) (#1355)

2026-08-04 13:52:18 +05:30

test_gpu_worker_sizing.py

fix(gpu): prevent the 8 GB-card OOM crash behind the "backend unreachable" wave (#567/#570/#571/#580+) (#586)

2026-06-21 04:51:45 +05:30

test_hf_cache_fallback.py

fix(windows): HF cache disk-fallback for WinError 448 (plan-01, closes #117 #118) (#137)

2026-05-29 09:27:43 +05:30

test_hf_cache_repair.py

fix(security): keep private diagnostics out of API responses (#1454)

2026-08-10 12:01:12 +00:00

test_hf_mirror_error_class.py

test(security): validate pinned repair fixtures

2026-08-10 03:49:23 +00:00

test_hf_mirror_settings.py

fix: first-run wizard version + mirror-unreachable rescue + lifecycle-aware backend reachability (#1094)

2026-07-12 11:05:00 +05:30

test_hf_revisions.py

fix(security): pin NLLB model revision

2026-08-10 07:27:43 +00:00

test_identity_paths_survive_the_rename.py

fix: restore generation and startup broken by the rename (#1417) (#1420)

2026-08-08 02:40:52 +05:30

test_install_disk_space.py

fix(security): pin curated Hugging Face revisions

2026-08-09 21:07:01 +00:00

test_issue_fixes.py

chore(repo): declutter root — move community docs into .github/

2026-07-21 13:32:13 +05:30

test_job_queue.py

feat: add audio trimming for reference clips, implement streaming transcription, and refactor ffmpeg utility handling

2026-04-18 22:37:20 +05:30

test_job_store.py

refactor: update backend architecture, expand frontend state management, and synchronize voice-pro research modules.

2026-04-21 18:32:25 +05:30

test_kitten_input_shape_1173.py

fix: resolve open issue batch — #1172 #1173 #1174 #1185 #1186 #1188

2026-07-19 04:29:33 +05:30

test_llm_endpoint_settings.py

test: make LLM-provider state leaks between tests impossible (#878) (#894)

2026-07-02 19:06:14 +05:30

test_llm_openai_available.py

fix(deps): bundle openai so Cinematic/LLM features work out of the box (#484)

2026-06-16 14:36:59 +05:30

test_llm_providers_router.py

fix(security): stabilize streamed failure responses

2026-08-09 21:50:42 +00:00

test_llm_providers.py

test(llm): complete OrcaRouter provider contract

2026-08-11 16:47:43 +00:00

test_llm_skills.py

fix(security): stabilize streamed failure responses

2026-08-09 21:50:42 +00:00

test_llm_state_isolation.py

test: make LLM-provider state leaks between tests impossible (#878) (#894)

2026-07-02 19:06:14 +05:30

test_lmstudio_model_discovery.py

fix(llm): LM Studio translation failed on a placeholder model name (#1332) (#1359)

2026-08-04 14:48:56 +05:30

test_load_budget_split_1033.py

test: stop the load-budget test racing the AudioSeal cold start (#1363)

2026-08-04 17:19:45 +05:30

test_local_stats.py

feat(privacy): Settings → Usage — local-only insights instead of cloud analytics (#1114)

2026-07-12 17:15:10 +05:30

test_locale_parity.py

fix(i18n): cover Arabic dub history plurals

2026-08-11 14:32:05 +00:00

test_logging_utils.py

fix: tighten log safety regressions

2026-08-10 12:49:42 +00:00

test_longform_e2e.py

fix(audiobook): say why a chapter failed, and don't hang when an engine stops (#1321) (#1325)

2026-07-30 04:22:50 -07:00

test_longform_import.py

fix(audiobook): add AudiobookPlan.chapter_count so import doesn't 500 (#544)

2026-06-20 09:15:40 +05:30

test_longform_jobs.py

fix(longform): bind job_store at import so the library route can't resolve a leaked module world

2026-07-16 21:10:25 +05:30

test_longform_limits.py

fix(longform): correctness + robustness fixes from adversarial review (#418)

2026-06-13 16:31:13 +05:30

test_longform_parser.py

feat(longform): canonical Python parser + golden corpus (#27 slice A) (#465)

2026-06-14 18:26:17 +05:30

test_longform_render.py

feat(longform): two-pass loudnorm builders + parser (#28 slice 1 — pure) (#449)

2026-06-14 13:10:01 +05:30

test_longform_resume.py

feat(audiobook): durable crash-resume for interrupted longform renders (#470)

2026-06-14 23:30:11 +05:30

test_longform_segment_cache.py

feat(longform): segment-level render cache — edit one sentence, re-render one segment (#1048)

2026-07-10 17:40:15 +05:30

test_longform_synth_params.py

fix(tts): Vietnamese consistency — Voice vs Audiobook divergences (#1142)

2026-07-13 23:42:36 +05:30

test_loopback_server_mode.py

Merge remote-tracking branch 'origin/main' into fix/ghas-path-boundary

2026-08-10 04:32:23 +00:00

test_loudness.py

feat(longform): two-pass loudnorm measure orchestrator + wiring (#28 slice 2) (#455)

2026-06-14 16:03:28 +05:30

test_low_vram_advisory.py

fix(routing): #1226 review — scope the VRAM floor to where it was measured

2026-07-23 03:28:22 +05:30

test_mastering_no_hidden_reverb.py

fix(audio): remove hidden reverb from the mastering pre-stage — reverb is preset-declared only (#986)

2026-07-08 02:26:19 +05:30

test_mcp_bindings.py

feat: TTS-only first run, platform-curated ASR, guided OS permissions, parakeet-mlx

2026-07-17 15:20:54 +05:30

test_mcp_graceful_degradation.py

fix(backend,shell): a missing MCP SDK can no longer kill the backend — and a failed setup now self-heals (#1156)

2026-07-16 01:45:00 +05:30

test_mcp_mount.py

feat(mcp): OMNIVOICE_MCP_ALLOWED_HOSTS — configurable host allowlist for MCP transport security (#1249) (#1250)

2026-07-26 02:02:23 -07:00

test_mcp_startup_timeout.py

fix(startup): timeout-bound MCP session-manager start to stop M1 startup hang (#632) (#645)

2026-06-23 12:49:39 +05:30

test_media_tool_missing.py

fix(errors): six failures that reached users as raw text (#1262, #1256, #1251, #1247, #1257, #1254) (#1264)

2026-07-26 12:39:29 -07:00

test_media_tools.py

fix(security): move host-path authority into the desktop shell (#1448)

2026-08-09 21:55:43 +00:00

test_memory_budget.py

feat(memory): honest /model/loaded accounting + a free-memory budget probe (#1111)

2026-07-12 16:11:42 +05:30

test_migration_0006_instruct.py

fix(voice): stop poisoning design profiles with "[object Object]" instruct (#556)

2026-06-20 08:50:42 +05:30

test_migration_0007_instruct_rebuild.py

fix(design): heal validator-rejecting instruct on design voices (#594/#571/#596) (#600)

2026-06-22 05:21:19 +05:30

test_missing_clone_ref_is_not_silent.py

fix(dub): say when a clone reference is gone instead of rendering a default voice (#1331) (#1361)

2026-08-04 16:56:52 +05:30

test_mlx_audio_instruct_1405.py

fix(mlx): pass the voice description the design model requires (#1405)

2026-08-07 16:15:31 +05:30

test_mlx_gate_390.py

fix(engines): MLX platform gate (#390) + ASR gpu_compat + IndexTTS2 (#21 PR 2/5) (#431)

2026-06-14 01:43:21 +05:30

test_mlx_supported.py

feat(routing): host device probe + routing resolver (#21 PR 1/5) (#430)

2026-06-13 23:03:22 +05:30

test_mm2_lifecycle.py

fix(memory): release the model before emptying the cache, not after

2026-08-12 02:33:04 +05:30

test_model_cache_repair.py

fix: repair the resolved Hugging Face cache

2026-08-10 07:04:42 +00:00

test_model_checkpoint_resolve.py

feat: TTS-only first run, platform-curated ASR, guided OS permissions, parakeet-mlx

2026-07-17 15:20:54 +05:30

test_model_install_dedup.py

fix(desktop): make recording and dubbing reliable (#1481)

2026-08-10 22:43:37 +00:00

test_model_load_extends_generate_budget.py

fix(generate): a healthy model download extends the budget it was blowing (#1367) (#1379)

2026-08-05 18:11:58 +05:30

test_model_load_timeout.py

fix(backend): self-healing GPU pool so a reset can't strand requests (#589/#599) (#601)

2026-06-22 04:42:16 +05:30

test_model_lock_event_loop_1417.py

fix(tts): a cold model load from a GPU-pool worker no longer waits on the wrong event loop (#1417) (#1418)

2026-08-08 03:02:10 +05:30

test_model_manager_preload.py

fix(models): repair corrupt configs and isolate ASR preload (#1437)

2026-08-09 20:24:25 +00:00

test_models_catalog.py

feat: TTS-only first run, platform-curated ASR, guided OS permissions, parakeet-mlx

2026-07-17 15:20:54 +05:30

test_models_dir_setting.py

Merge remote-tracking branch 'origin/main' into fix/ghas-path-boundary

2026-08-10 04:32:23 +00:00

test_moss_entry_point_1287.py

fix(engines): verify MOSS-TTS-Nano's entry point, don't just import the module (#1316)

2026-07-29 16:41:16 -07:00

test_moss_tts_v15.py

test(security): validate pinned repair fixtures

2026-08-10 03:49:23 +00:00

test_nemo_parakeet_install_hint.py

fix(engines): nemo-parakeet install hint stops recommending a shared-venv-breaking pip install (#974) (#991)

2026-07-08 04:26:12 +05:30

test_network_beats_install_diagnosis.py

fix(errors): stop giving advice that cannot work (#1347, #1335, #1334) (#1374)

2026-08-05 14:45:29 +05:30

test_network_middleware.py

polish(network): outermost PIN gate + non-buffering ASGI middleware + listener test (#160)

2026-05-30 11:45:54 +05:30

test_network_share_lifecycle.py

polish(network): outermost PIN gate + non-buffering ASGI middleware + listener test (#160)

2026-05-30 11:45:54 +05:30

test_network_share.py

fix: centralize LAN share runtime state

2026-08-10 09:29:42 +00:00

test_no_committed_analytics_token.py

feat(analytics): in-repo publishable token — source builds get the same consent-gated analytics (#1193)

2026-07-20 11:51:40 +05:30

test_no_conflict_markers.py

test: fail CI on unresolved git conflict markers in tracked files

2026-07-20 15:58:34 +05:30

test_no_hardcoded_cjk.py

fix(errors): six failures that reached users as raw text (#1262, #1256, #1251, #1247, #1257, #1254) (#1264)

2026-07-26 12:39:29 -07:00

test_no_literal_borders.py

fix(theme): remove stray token-border frames + make accent family theme-track (#864)

2026-07-01 22:06:56 +05:30

test_no_orphan_service_modules.py

chore(backend): delete dead services/batched_tts.py + guard against orphaned service modules (#1086)

2026-07-12 02:17:07 +05:30

test_no_unspeakable_chunks_1330.py

fix(generate): never send a chunk with nothing to say (#1330)

2026-08-07 01:30:24 +05:30

test_normalize_audio_silence.py

fix(audio): stop near-silent renders becoming blank noise + guard archetype renders (#204)

2026-05-31 13:27:30 +05:30

test_omnivoice_lazy_model_import.py

fix(startup): don't let one optional transformers symbol kill the backend (#1229)

2026-07-23 00:28:42 +05:30

test_omnivoice_path.py

fix(bootstrap): gate venv on omnivoice import + source fallback (#564) (#603)

2026-06-22 05:07:07 +05:30

test_onboarding_demo.py

test(onboarding): guard demo clip stays un-ignored + a bundled resource (#621) (#648)

2026-06-23 16:57:48 +05:30

test_onset_align.py

fix(dub): dialogue starts stop snapping to footsteps — sustained-energy onsets, bounded snap (#963) (#967)

2026-07-05 21:15:14 +05:30

test_openai_speech_binary_guard_1172.py

fix: resolve open issue batch — #1172 #1173 #1174 #1185 #1186 #1188

2026-07-19 04:29:33 +05:30

test_pause_markers.py

feat(tts): inline [pause Nms] marker for silence in generated speech (#276) (#277)

2026-06-04 08:01:29 +05:30

test_persona_bundle.py

feat(persona): .ovsvoice build/parse core + embed_watermark(force=) (#29 slice A) (#460)

2026-06-14 17:55:17 +05:30

test_persona_kind.py

fix(persona): preserve design kind + vd_states across share/import (Wave 5 §R3) (#405)

2026-06-13 02:21:43 +05:30

test_phase4_services.py

refactor: update backend architecture, expand frontend state management, and synchronize voice-pro research modules.

2026-04-21 18:32:25 +05:30

test_pitch_stretch_async.py

fix(dub): async-ify _pitch_preserving_stretch (#133 Greptile P1) (#152)

2026-05-29 23:13:32 +05:30

test_pkg_resources_available.py

fix(asr): cap setuptools <80 so pkg_resources stays present (#224) (#225)

2026-06-01 11:46:32 +05:30

test_pockettts_deferred.py

test(ci): scan both workflow extensions

2026-08-10 12:50:00 +00:00

test_pockettts_sidecar.py

fix(pockettts): close license review findings

2026-08-10 09:24:37 +00:00

test_port_in_use_exit.py

fix(startup): explain a lost port race instead of exiting 1 in silence (#1364) (#1370)

2026-08-05 02:14:27 +05:30

test_prefs_thread_safety.py

feat(engines): one-click IndexTTS-2 sidecar install from Settings → Engines (#1083)

2026-07-12 01:13:35 +05:30

test_profile_consent.py

feat(profiles): consent-locked voice profiles — verified_own_voice + spoken consent flow (Wave 0.2) (#354)

2026-06-11 22:10:30 +05:30

test_profile_design_save_decouple.py

fix(clone): voice-design panel no longer crashes on a partial vd_states shape (#983) (#995)

2026-07-08 04:34:14 +05:30

test_profile_language_propagation.py

fix(lang): propagate profile/request language into generation + longform (#533/#505/#502) (#565)

2026-06-20 14:45:58 +05:30

test_profile_unification.py

fix(clone): voice-design panel no longer crashes on a partial vd_states shape (#983) (#995)

2026-07-08 04:34:14 +05:30

test_pronunciation_api.py

fix(test): close the reload-induced test-isolation leak at its source (#932 follow-up) (#933)

2026-07-03 18:07:44 +05:30

test_pronunciation.py

feat(tts): user pronunciation dictionary (expressive-tts slice 1) (#685)

2026-06-25 03:16:25 +05:30

test_published_image_path_is_pinned.py

chore: point every repo reference at github.com/debpalash/VoiceStudio (#1394)

2026-08-06 18:15:56 +05:30

test_pyannote_hf_compat.py

fix(diarization): use_auth_token->token shim for pyannote on HF Hub 1.x (#167) (#168)

2026-05-30 17:11:44 +05:30

test_pytorch_whisper_fallback.py

fix(desktop): make recording and dubbing reliable (#1481)

2026-08-10 22:43:37 +00:00

test_pyvideotrans_contract.py

test(api): pin the /generate surface pyvideotrans's integration consumes (Wave 1.3) (#359)

2026-06-11 23:06:49 +05:30

test_recurrence_hardening.py

fix(engines): bundle en_core_web_sm — no mid-generation GitHub download (#1146)

2026-07-14 00:48:23 +05:30

test_redub_incremental.py

fix(watermark): route every synthetic-audio producer through one mark_synthetic chokepoint (#1169)

2026-07-16 23:08:36 +05:30

test_release_preview_asset_cleanup.py

ci(release): stop the macOS preview updater bundle colliding with itself (#1281)

2026-07-29 01:16:47 -07:00

test_release_preview_manifest_rebuild.py

feat: rename the product to VoiceStudio (previously OmniVoice-Studio)

2026-08-07 01:30:58 +05:30

test_resolve_heartbeat_1414.py

test: always release parked heartbeat writer

2026-08-09 22:42:48 +00:00

test_resplit_speaker.py

fix(dub): speaker-aware re-split so merged speaker turns separate (#486) (#616)

2026-06-22 22:39:07 +05:30

test_response_safety.py

fix(security): keep private diagnostics out of API responses (#1454)

2026-08-10 12:01:12 +00:00

test_rocm_arch_gate.py

fix(rocm): #1228 review — a pre-set HSA override must also name a real target

2026-07-23 04:24:20 +05:30

test_rocm_torch_pins_match_pyproject.py

fix(colab): pin torchvision==0.23.0 to resolve torchvision::nms operator missing error (#1357) (#1358)

2026-08-04 16:27:36 +05:30

test_router_smoke.py

fix(test): router-smoke tests leak-proof against full-suite order (#932)

2026-07-03 17:47:45 +05:30

test_routing_redaction.py

fix(engines): revive dead license dialog, refresh matrix on select, surface routing verdict (#905)

2026-07-02 23:22:49 +05:30

test_scrub.py

fix(security): fail closed on cleanup and redaction errors

2026-08-09 21:33:45 +00:00

test_segment_refs.py

feat(dub): per-segment clone references (Wave 3.2) (#369)

2026-06-12 12:50:35 +05:30

test_segmentation.py

fix(dub): the speaker-count hint is honored on every diarization path + clone-purity guard (#952)

2026-07-04 22:53:47 +05:30

test_segmented_download_default.py

feat(setup): faster downloads by default + prominent, encouraged HF-token entry (#669)

2026-06-24 12:35:57 +05:30

test_sentence_chunker.py

feat(stream): sentence-by-sentence /ws/tts via ported chunker (Wave 1.4) (#358)

2026-06-11 23:04:00 +05:30

test_settings_updates_api.py

feat(updates): data-safe updates — pre-migration DB backups, guarded venv heal, release notes + changelog reader (#909)

2026-07-02 23:52:17 +05:30

test_setup_preflight.py

feat(network): automatic Hugging Face endpoint selection — probe, pick, remember (#1082)

2026-07-11 23:44:21 +05:30

test_setup_recommendations.py

feat(workers): pin work to the chosen GPU, and say when its model is missing

2026-08-11 15:24:11 +05:30

test_sherpa_dictation.py

fix(security): preserve authorized legacy snapshots

2026-08-09 21:14:12 +00:00

test_sherpa_model_sizes.py

fix(dictation): stop the widget stranding an empty square, repair the swept data dirs (#1398)

2026-08-07 05:39:41 +05:30

test_sherpa_ws_streaming.py

feat: TTS-only first run, platform-curated ASR, guided OS permissions, parakeet-mlx

2026-07-17 15:20:54 +05:30

test_shutdown_preload_race_1000.py

fix(backend): shutdown wait bound 3s→20s — post-merge review finding on #1002; absorb #1015's design-path test (#1020)

2026-07-09 02:10:06 +05:30

test_sidecar_downmix_is_channel_aware.py

fix(engines): downmix along the channel axis, not axis 0 (#1328) (#1366)

2026-08-04 23:50:06 +05:30

test_sidecar_install.py

feat(indextts): add native IndexTTS 2.5 support (#1485)

2026-08-11 01:26:25 +00:00

test_sidecar_stdout_isolation_1428.py

fix(engines): give every sidecar a private fd for its frames (#1428)

2026-08-09 20:22:06 +00:00

test_smart_fit_export.py

fix(security): bind native paths to picker capabilities

2026-08-10 04:37:45 +00:00

test_smart_fit_generate.py

fix(watermark): route every synthetic-audio producer through one mark_synthetic chokepoint (#1169)

2026-07-16 23:08:36 +05:30

test_socks_proxy.py

fix(security): keep private diagnostics out of API responses (#1454)

2026-08-10 12:01:12 +00:00

test_sonitranslate_venv.py

fix(win): SoniTranslate venv paths cross-platform — Scripts/ on Windows (#186) (#188)

2026-05-31 03:38:54 +05:30

test_spa_inject.py

fix(docker): runtime API-base override so served deployments reach the backend (#174)

2026-05-30 20:54:49 +05:30

test_speaker_clone_purity.py

fix(dub): restore cast voices from source audio (#1484)

2026-08-10 23:46:37 +00:00

test_speaker_hint.py

feat: TTS-only first run, platform-curated ASR, guided OS permissions, parakeet-mlx

2026-07-17 15:20:54 +05:30

test_srt_parser.py

feat(dub): paste a translation from an external source onto existing segments

2026-07-20 13:18:51 +05:30

test_ssml_lite.py

feat(longform): SSML-lite prosody markup — [slow]/[fast]/[emphasis]/[spell] (PR 8b) (#421)

2026-06-13 16:56:55 +05:30

test_storage_report.py

fix(uninstall,storage): remove the saved-env leftover; count sidecar engines in disk usage (#1108)

2026-07-12 15:39:07 +05:30

test_stories_encode.py

feat(stories): MP3 export — backend ffmpeg encode + format selector (#181)

2026-05-30 23:09:43 +05:30

test_subprocess_fallback.py

fix(win): subprocess spawns work under bun run dev (--reload) on Windows (#122) (#175)

2026-05-30 21:15:09 +05:30

test_subprocess_sidecar_wire.py

test(engines): parametrize wire-protocol tests across all sidecar engines (#1408)

2026-08-07 17:53:15 +05:30

test_subtitle_segmenter.py

refactor: update backend architecture, expand frontend state management, and synchronize voice-pro research modules.

2026-04-21 18:32:25 +05:30

test_supertonic3.py

fix(ci): skip supertonic license/cpu tests when the optional dep is absent (#211)

2026-06-01 08:12:55 +05:30

test_synth_error_classes.py

fix(synth): #1227/#1221 review — assert the marker through classify()

2026-07-23 04:24:54 +05:30

test_synthetic_audio_watermark_1169.py

feat(audiobook): cast/voice mapping + markup toolbar + live stats + validation (#1217)

2026-07-21 01:54:13 +05:30

test_tailscale_service.py

feat(network): user-configurable backend / LAN-share / UI ports (#163)

2026-05-30 12:52:51 +05:30

test_text_normalization_routes.py

feat(tts): normalization covers the OpenAI-compat API, streaming, and batch paths (#1054)

2026-07-11 00:05:04 +05:30

test_text_normalization.py

fix(security): make control filtering boundaries explicit

2026-08-10 07:30:08 +00:00

test_text_polish.py

feat(dictation): rebuild to Wispr-Flow quality — live waveform, streaming commits, honest insertion, polished text (#888)

2026-07-02 17:57:56 +05:30

test_tls_dropped_classification.py

fix(errors): explain a cut TLS connection instead of printing _ssl.c:1016 (#1312)

2026-07-29 14:17:40 -07:00

test_torch_compile_gate.py

fix(tts): torch.compile failures fall back to eager — generation never fails on unsupported GPUs (#278) (#327)

2026-06-11 13:09:18 +05:30

test_torch_compile_path_gate.py

test: reset model-manager shutdown state between backend tests (#1320)

2026-07-30 01:30:06 -07:00

test_torch_constraints_are_applied.py

Merge main and address Docker review findings

2026-08-10 04:53:37 +00:00

test_torch_dtype_guard.py

test(ci): root-cause and neutralize the flaky trio — a leaked torch fp16 default dtype (#1021)

2026-07-09 02:19:07 +05:30

test_transcribe_reference.py

feat: TTS-only first run, platform-curated ASR, guided OS permissions, parakeet-mlx

2026-07-17 15:20:54 +05:30

test_translate_degraded_not_failed.py

fix(dub): a rate-limited polish pass no longer skips fitting, fails the UI, or ignores Retry-After (#1135)

2026-07-13 16:13:41 +05:30

test_translation_engine_install.py

fix(engines): pin uv pip install to the running interpreter (#529/#527) (#559)

2026-06-20 09:11:20 +05:30

test_translation_install_command.py

feat(translate): highlighted Install affordance for uninstalled engines + dismissable/auto-clearing error banner (#847)

2026-07-01 11:16:13 +05:30

test_translator.py

test(translator): stop patching time.sleep for the whole process (#1390)

2026-08-06 14:29:14 +05:30

test_truncated_download_retry.py

fix(download): #1224 review — a settled Hub verdict must not retry

2026-07-23 04:29:17 +05:30

test_truthful_degraded_state.py

fix(memory): release the model before emptying the cache, not after

2026-08-12 02:33:04 +05:30

test_tts_placement_self_heal.py

test(#1191): make the self-heal restore deterministic (wait in-loop, not fire-and-forget race)

2026-07-21 03:24:25 +05:30

test_unified_memory_offload.py

fix(memory): on unified memory, "offload" must mean UNLOAD — the 16 GB dub OOM (#1119) (#1122)

2026-07-12 18:33:36 +05:30

test_unload_advertised_ids.py

fix(errors): six failures that reached users as raw text (#1262, #1256, #1251, #1247, #1257, #1254) (#1264)

2026-07-26 12:39:29 -07:00

test_user_env_authoritative.py

feat(hardening): six recurrence guards from the closed-issue-history audit (#1141)

2026-07-13 23:19:58 +05:30

test_user_env.py

feat(settings): configurable models directory (#64) (#149)

2026-05-29 16:52:50 +05:30

test_uv_cross_drive.py

fix: bot-review harvest (16 findings) + deterministic style/locale CI + reviewer configs

2026-07-20 09:47:41 +05:30

test_vendored_omnivoice_imports_1417.py

fix(tts): a cold model load from a GPU-pool worker no longer waits on the wrong event loop (#1417) (#1418)

2026-08-08 03:02:10 +05:30

test_video_context.py

test(video): enforce exact Pillow floor

2026-08-10 02:18:02 +00:00

test_voice_design_instruct_heal.py

fix(security): align bounded model validation

2026-08-10 00:04:53 +00:00

test_voicestudio_branding.py

fix: restore pre-release version 0.4.2

2026-08-11 04:28:38 +00:00

test_voxcpm2_guardrails.py

fix(tts): voxcpm2 — version floor, reference-clip prep, trailing-silence guard (#1055)

2026-07-11 00:21:54 +05:30

test_watermark_chunking_1045.py

fix(memory): release the model before emptying the cache, not after

2026-08-12 02:33:04 +05:30

test_watermark_route_coverage.py

feat(workers): make a remote GPU actually run a task, end to end

2026-08-11 07:16:04 +05:30

test_wedge_vs_slow_hardware_1416.py

fix(generate): a job that never ran was not too heavy for your computer (#1416) (#1424)

2026-08-08 15:03:51 +05:30

test_why_no_gpu_1274.py

feat(diagnostics): say why a GPU host fell back to CPU (#1274) (#1425)

2026-08-08 15:23:38 +05:30

test_window_launch_state.py

fix(desktop): make recording and dubbing reliable (#1481)

2026-08-10 22:43:37 +00:00

test_windows_paging_file.py

fix(errors): six failures that reached users as raw text (#1262, #1256, #1251, #1247, #1257, #1254) (#1264)

2026-07-26 12:39:29 -07:00

test_windows_stdio_guards.py

fix(win): stop the Fortran-runtime console-close abort and cp1252 UnicodeEncodeError crash classes (#1153, #1155)

2026-07-16 01:45:23 +05:30

test_worker_agent.py

feat(workers): pin work to the chosen GPU, and say when its model is missing

2026-08-11 15:24:11 +05:30

test_worker_breaker.py

feat(workers): remote GPU workers over a versioned gRPC protocol

2026-08-10 14:18:42 +05:30

test_worker_capabilities_downloaded.py

feat(workers): dubbing goes remote, and the protocol stops lying to old workers

2026-08-11 17:39:46 +05:30

test_worker_capability_refresh.py

feat(workers): remote downloads, audiobook chapters, and one port that stays honest

2026-08-11 16:53:33 +05:30

test_worker_capacity.py

feat(workers): pin work to the chosen GPU, and say when its model is missing

2026-08-11 15:24:11 +05:30

test_worker_client_liveness.py

feat(workers): one gateway for GPU calls, and results too big for the wire

2026-08-11 13:30:26 +05:30

test_worker_deadlines.py

feat(workers): dubbing goes remote, and the protocol stops lying to old workers

2026-08-11 17:39:46 +05:30

test_worker_errors.py

feat(workers): dubbing goes remote, and the protocol stops lying to old workers

2026-08-11 17:39:46 +05:30

test_worker_executor_residency.py

feat(workers): one gateway for GPU calls, and results too big for the wire

2026-08-11 13:30:26 +05:30

test_worker_identity.py

feat(workers): remote GPU workers over a versioned gRPC protocol

2026-08-10 14:18:42 +05:30

test_worker_inbound_keys.py

fix(models): do not preload a model on a machine with no local user

2026-08-12 01:40:22 +05:30

test_worker_inbound_transport.py

fix(workers): unload idle models on an inbound-only node

2026-08-12 01:14:16 +05:30

test_worker_inputs.py

feat(workers): one gateway for GPU calls, and results too big for the wire

2026-08-11 13:30:26 +05:30

test_worker_lifecycle.py

feat(workers): make a remote GPU actually run a task, end to end

2026-08-11 07:16:04 +05:30

test_worker_protocol_contract.py

feat(workers): dubbing goes remote, and the protocol stops lying to old workers

2026-08-11 17:39:46 +05:30

test_worker_protocol_gen.py

feat(workers): remote downloads, audiobook chapters, and one port that stays honest

2026-08-11 16:53:33 +05:30

test_worker_registry.py

feat(workers): remote GPU workers over a versioned gRPC protocol

2026-08-10 14:18:42 +05:30

test_worker_routing.py

feat(workers): dubbing goes remote, and the protocol stops lying to old workers

2026-08-11 17:39:46 +05:30

test_worker_scheduler.py

feat(workers): dubbing goes remote, and the protocol stops lying to old workers

2026-08-11 17:39:46 +05:30

test_worker_server_integrity.py

feat(workers): dubbing goes remote, and the protocol stops lying to old workers

2026-08-11 17:39:46 +05:30

test_worker_service_api.py

feat(workers): dubbing goes remote, and the protocol stops lying to old workers

2026-08-11 17:39:46 +05:30

test_worker_task_store.py

feat(workers): pin work to the chosen GPU, and say when its model is missing

2026-08-11 15:24:11 +05:30

test_worker_transport.py

feat(workers): dubbing goes remote, and the protocol stops lying to old workers

2026-08-11 17:39:46 +05:30

test_worker_upload_client.py

feat(workers): one gateway for GPU calls, and results too big for the wire

2026-08-11 13:30:26 +05:30

test_worker_upload_server.py

feat(workers): dubbing goes remote, and the protocol stops lying to old workers

2026-08-11 17:39:46 +05:30

test_youtube_cookie_import.py

fix: harden cookie import lifecycle

2026-08-10 04:01:44 +00:00

test_yt_drm_escalation.py

fix(errors): six failures that reached users as raw text (#1262, #1256, #1251, #1247, #1257, #1254) (#1264)

2026-07-26 12:39:29 -07:00