ASR head-to-head for video transcription: ByteDance Doubao wins with millisecond timestamps
vista8 · x · 2026-10-05
Testing cloud ASR services for transcribing subtitle-less YouTube and Bilibili videos, the author compared Doubao Speech (turbo), Qwen3-ASR on SiliconFlow, GLM-ASR-2512, Step 2.5 ASR, and Xiaomi MiMo-ASR. Doubao came out best with millisecond-level timestamps and fast speed. These features are coming to the Qiaomu clipping plugin, adding local Whisper and local Qwen3 ASR alongside cloud speech-to-text APIs.
More from Apps
- Obsidian Starter Kit turns meeting decisions into queryable note properties — dSebastien · 2026-10-06
- AI tools quietly ate 44 GB of his disk, so he built free cleaner Sparewise — PossibilityKind3028 · 2026-10-06
- Claude Projects cloud sessions can now read and edit approved local folders in place — trq212 · 2026-10-06
- Claude can now join Slack group DMs and follow threads like a teammate — maxleiter · 2026-10-06
- FlattenSF: A Tool That Finds the Flattest Route Between Any Two Points in San Francisco — ishan0102 · 2026-10-06
- Family-run lender built its own tool on Replit, grew business 15% in 3 months — amasad · 2026-10-06