18.4k stars · Apache-2.0 · v3.0.1 (2026-02-28) · Track this in Scout
One-click transcribe, translate and dub pipeline; doubles every video into a Russian edition.
▶Repo detailsthe review · specs · pros & cons
What it is
A single pipeline that transcribes a video with Whisper, splits the subtitles at natural sentence boundaries, translates them through an LLM using a translate-then-critique-then-adapt pass, and optionally dubs the result in a synthetic or cloned voice that stays time-aligned with the picture.
What it is good for. This is the machine that turns a bilingual publishing plan into a system: one Camtasia export in, an English edition and a Russian edition out. Doubling a video catalogue without doubling filming time is a cheap growth lever — and the Russian-language SaaS tutorial space is far less crowded than the English one. The SRT files it produces are also free blog copy and free chapter markers.
- Handles the whole chain, not one step of it
- The critique pass makes translations markedly better than raw machine output
- Speaks to ElevenLabs, Azure and edge-tts — it slots into services many people already pay for
- Python/conda install with a Streamlit UI; dependency conflicts are common
- Local Whisper really wants a GPU — on a laptop or a plain CPU server you're pushed to cloud ASR
- Dubbed audio drifts slightly in length, so screencasts with tight on-screen callouts still need a manual pass
