Upload and automated pipeline
After source upload, the console shows audio extraction, vocal split, ASR, speaker routing, literal reference, AI adaptation, and TTS timing.
Review console
Review like an editing suite, not after export.
The web console is not about manually choosing models. It turns backend automation into a script editing surface that humans can understand, adjust, and approve.
Project: Edge of the Stars Ep.01
Processing CompleteAll saved locally
We don't have much time before they lock down the sector.
00:12:34:08 / 00:43:12:15
1.0xCC
00:12:2000:12:3000:12:4000:12:5000:13:00
Speaker Recognition
L
Female / 2896%Calm AssuredLena Voss
SPK_01
J
Male / 3194%Deep SteadyJian Hao
SPK_02
M
Male / 4592%Gravelly AuthoritativeMarcus Hale
SPK_03
I
Female / 3491%Warm IntelligentIra Solari
SPK_04
Quality Overview
1016
Ready64263.2%
Review15615.4%
Fix (Must Fix)686.7%
Untimed424.1%
AI Suggested10810.6%
| line_id | Speaker | Start | End | Original / ASR | Literal Reference | Chinese Dub Script (Editable) | Emotion | TTS Duration | Slot | Collision Risk | Status | Actions |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 130 | 00:01:20:12 | 00:01:24:12 | We don't have much time before they lock down the sector. | 我们在他们封锁这片区域之前，只剩下不多的时间了。 | 2.48s | 2.40s | Low | ready | ||||
| 131 | 00:01:24:12 | 00:01:26:24 | Then we move now. | 那我们现在就行动。 | 1.92s | 2.12s | Medium | review | ||||
| 132 | 00:01:26:24 | 00:01:28:18 | Understood. | 明白。 | 0.98s | 1.54s | Low | ready | ||||
| 133 | 00:01:28:18 | 00:01:32:02 | I'll prep the jump drive and plot a vector. | 我会准备跃迁引擎并规划路线。 | 3.28s | 3.60s | High | fix | ||||
| 134 | 00:01:32:02 | 00:01:34:15 | How long to charge? | 充能需要多久？ | 1.78s | 2.13s | Medium | review | ||||
| 135 | 00:01:34:15 | 00:01:37:10 | Thirty seconds. Maybe less with overdrive. | 三十秒，超频的话可能更短。 | 2.52s | 2.88s | Low | ai | ||||
| 136 | 00:01:37:10 | 00:01:40:01 | Make it quick. | 动作要快。 | — | 2.60s | — | untimed |
Editor-style preview
The center area combines video preview, subtitle overlay, and multi-track timeline. Selecting a line syncs preview and quality report.
References plus final input
Each row keeps source/ASR, literal reference, AI adaptation, and final editable dubbing text. Humans mainly edit the final line.
Live color quality
Rows show green, yellow, red, blue, or gray based on text length, real TTS duration, slot length, spare ms, collision, and broken-line risk.