Start with your video
Work with a local video or prepare a YouTube source using the project’s tools. English to French is this adaptation’s primary workflow.
LangSwapVideo · English → French
Another language.
The same story.
Turn an English video into a French version: transcription, translation, voice and subtitles brought together in one workflow.
Project in development · Ask about availability
ENEnglish source
FRFrench version
Speaker 01 · 00:18 — 00:22
ENLet’s take a moment to explore.
FRPrenons un moment pour découvrir.
French video + source SRT + French SRT
01Transcribe and translate02Adapt the voices03Export video and subtitles
01 / From source to French version
A complete workflow.
A video to return to.
LangSwap brings the steps of preparing a French version together, from the first recognized word to the output files.
Recover the words
Transcription splits speech into timestamped segments. Speaker detection helps distinguish people throughout the video.
Move into French
Translation prepares French text for each segment. Wording, proper names and important passages still need review before distribution.
A reference for each voice
The adapted profile uses a reference per speaker to aim for a consistent timbre when they speak again. Resemblance remains an approximation to check by listening.
Put speech back in time
Audio segments are placed against the video’s time markers. The workflow also lets you revisit text, timing or speaker assignments in the editor.
Get useful output files
Export the French video and two SRT files: source transcription and translation. Keep a version to watch and subtitles you can revise.
02 / Your workflow
From English to French,
step by step.
Prepare
Select the video and processing profile.
Transcribe
Identify speech, timestamps and speakers.
Translate and dub
Prepare French text and voice segments.
Review and export
Check the result, make corrections and collect the files.
03 / A local working profile
Your video.
Your environment.
The documented French profile runs on a machine with the required models: speech recognition, translation and voice generation. This workflow requires no API key and does not send audio to a cloud service.
Installing the models may require Internet access. Other engines, including Edge voices, use online services when explicitly selected.
- Processing with installed models
- NVIDIA GPU acceleration where supported
- Project-specific voice and parameter choices
04 / In practice
A few things to know
Can I download LangSwap?
This page presents the project and its French adaptation. No purchase, public installer or online conversion service is offered here. Email me to discuss availability or a use case.
Which languages are targeted?
This adaptation’s verified workflow is English to French. Other engines appearing in the code do not mean every language and combination has been qualified.
Are the voices identical to the original?
No. The system aims for a consistent timbre per speaker, but resemblance, pronunciation and speaker separation may need corrections. A human listening review remains necessary.
What hardware do I need?
The documented French profile uses Linux or a suitable container, Python, FFmpeg and local models. NVIDIA/CUDA provides acceleration. The kit includes Windows launchers; it is not yet a ready-to-install consumer app.
Is the result automatically ready to publish?
Review the subtitles, listen to the voices and check names, numbers, translations and timing. Use content and voices you have permission to adapt.
A different way
to understand a video.
Have a question about the project or a video translation need? Get in touch.
Let’s talk about LangSwap ↗← All productsAdapted from the open-source project LangSwap.
Runtime licensed under AGPL-3.0-or-later; models and dependencies have their own licences.