One video. Seven languages.
Upload your video. Get it transcribed, translated and re voiced for Africa's audiences in minutes.
The problem
Made in one language, watched in many
African video is usually made in one language for an audience that speaks many. Adding a language has meant voice actors, studio time and weeks of work for each one, so most videos never get a second language at all.
520
living indigenous languages in Nigeria alone
Source: Ethnologue, Nigeria country profile872
films Nollywood produced in 2006, second only to India, all on video
Source: UNESCO Institute for Statistics, reported by UN News, May 2009200M+
people speak Kiswahili across East, Central and Southern Africa
Source: UNESCO, World Kiswahili Language Day65%
of the world's 396 million French speakers live in Africa
Source: OIF, La langue française dans le monde 2026, reported by Francopresse
How it works
Five quiet steps
Each step has a clear status in the app, and the whole run finishes with an email when your renders are ready.
Upload
Drop in an MP4 or MOV up to 30 minutes, or paste a link to a video you own.
Transcribe
Timestamps and speaker labels, synced to the player. Fix names and brand terms in place.
Translate
Every target language sits beside the source, line by line. Edit any line before it is voiced.
Voice
Pick a neural voice per language. The new track replaces the original sound and follows its timing.
Publish
Download dubbed videos, SRT and VTT files, and share review links with clients.
Upload a video
Drop in an MP4 or MOV up to 30 minutes, or paste a link to a video you own.
Links to videos you own work too
Uploads start after you sign in. Every render logs minutes, characters and storage to your usage meter.
How our AI works
The engine behind every dub
Every upload runs through these six stages, in this order, on the live pipeline. The card below them follows one real line of the sample episode through.
- 01
Speech recognition
A large speech model transcribes the original audio with a timestamp on every word and punctuation. The words are grouped into subtitle sized lines that break at sentence ends, at pauses of a second or more, or after about six seconds of speech.
Timed transcript and SRT
- 02
Line aligned translation
A large language model translates every line in place, never merging or dropping one, in a spoken register and at a similar length so it fits the same seconds on screen. Names, places, brands and numbers stay as they are. Yoruba keeps its tone marks and Igbo its dotted letters.
SRT and VTT per language
- 03
Neural voices
Each line is synthesized with the newest neural voice family available for the language, in the female or male voice picked for the job. Where no neural voice exists yet, the language is delivered as burned in subtitles and labeled that way.
One voice per language
- 04
Timing engine
Every voiced line is placed at the start time of the line it replaces. If a translation runs longer than its slot, it is sped up to close the gap, never beyond 1.45 times, so the speech stays natural. The finished track replaces the original audio.
Timing sync, not lip sync
- 05
Render
The new track is laid onto your original picture without re encoding whenever the codec allows, so the video quality you uploaded is the quality you get back. Subtitled languages are burned into a fresh encode.
MP4 per language
- 06
Quality check
Every finished line is checked against its slot. A voiced line that needed more than a 1.25 times speed-up, or still runs long at the 1.45 limit, is flagged, and so is any subtitle that reads faster than 20 characters a second. Flags show on the job with the timestamp, the line and the reason.
Flags per language
One real line, traced
- Heard at 00:00.0 · English
- You were 26 the first time somebody handed you a red envelope and you did the worst possible thing you could have done with it.
- Translated · French
- Tu avais 26 ans la première fois que quelqu'un t'a donné une enveloppe rouge et tu as fait la pire chose possible que tu aurais pu faire avec.
- Voiced with
- fr-FR-Chirp3-HD-Achird
One voice per language in each video for now. No lip sync and no voice cloning: timing sync only, stated plainly.
The workspace
What your team sees
Screens from the live workspace. The finished job is real: WellnessLab's Sciatica explainer, fetched from a link and dubbed into French, with the quality check the pipeline wrote for it.
Start a dub
Drop a file or paste a link, pick the languages and a voice, and list the names every language must keep as written.
Review what came back
Download every language, and see the lines worth a second look: each with its timestamp, its text and why it was flagged.
Watch it work
One real clip, dubbed
An explainer run through the live product: transcribed, translated, voiced in French and Swahili and captioned in Hausa. Every dubbed line in it is the pipeline's own output.
Coming next: the editor
Fix a name once, hear it right in every language
Next on our roadmap: the transcript synced to the video, with each target language beside the source for a line by line review before anything is voiced. Today you list the names and terms every language must keep as written, and every finished job flags the lines worth a second look.
Transcript
Timed to the videoTranslation review
French · voicedYou were 26 the first time somebody handed you a red envelope and you did the worst possible thing you could have done with it.
Tu avais 26 ans la première fois que quelqu'un t'a donné une enveloppe rouge et tu as fait la pire chose possible que tu aurais pu faire avec.
You opened it at the table, not in a corner, not later, not in the car on the way home.
Tu l'as ouverte à table, pas dans un coin, pas plus tard, pas dans la voiture en rentrant.
At the table in front of the man who had just given it to you with everybody watching and you looked at what was inside and said, "Thank you." in a slightly too loud voice.
À table, devant l'homme qui venait de te la donner devant tout le monde qui regardait et tu as regardé ce qu'il y avait dedans et tu as dit : « Merci » d'une voix un peu trop forte.
The room went quiet for about 2 seconds.
La pièce est devenue silencieuse pendant environ 2 secondes.
- Speaker labels
- Glossary lock
- Line by line review
- Timing check
Languages
Seven languages, one honest label each
Every language gets localized output. English, French and Swahili are voiced with neural voices. The other four start as subtitles and move to voice as neural voices become available, so nobody is surprised inside the app.
Voiced dubbing
Voiced dubbingFrançais
Voiced dubbingKiswahili
Subtitles now, voice coming
Subtitles now, voice comingYorùbá
Subtitles now, voice coming
Subtitles now, voice coming
Priced by the minute, because that is how it works
By the minute
Usage this month
Example workspace · Creator plan- Minutes uploaded42 of 60
- Minutes transcribed42
- Minutes translated84 across 2 languages
- Characters synthesized63,400
- Renders completed4
- Storage used1.8 GB
Every workspace logs minutes, characters and renders from the first upload. The same numbers are what your plan is measured against, per month.
For media houses and agencies
Built for teams with a back catalog
One workspace for the whole team, with the review tools an agency needs to hand a client seven versions of one film.
Team workspaces
Seats for editors, translators and reviewers, with roles, on Studio and Media House.
Client review links
Clients watch every language version and leave timestamped comments without an account.
Glossaries
Lock brand names, places and slang once. Every translation and every render respects them.
Back catalog volume
Bulk episode processing for series and archives is on the roadmap. Custom minutes are available today.
Founder case study
Built by a product engineer.
Agu builds web platforms for government and enterprise teams in Nigeria, and kept running into the same wall: one video that could reach ten times the audience, if only it spoke their language. Vocelle is that wall coming down. Here is an animated episode run through it: voiced in French and Swahili, captioned in Hausa, with its transcript and translation beside it.
Transcript
English (original)You were 26 the first time somebody handed you a red envelope and you did the worst possible thing you could have done with it.
You opened it at the table, not in a corner, not later, not in the car on the way home.
At the table in front of the man who had just given it to you with everybody watching and you looked at what was inside and said, "Thank you." in a slightly too loud voice.
Examples
Dubbed for WellnessLab
Scroll to turn the ring. WellnessLab's 3D health explainers, original plus two dubbed versions each.
Choose your plan
Creator
Creator
For YouTube and TikTok creators publishing weekly. Two languages, no watermark, per minute top ups when you need more.
- 60 minutes of source video a month
- Two target languages
- No watermark
- Per minute top ups
Prices in Naira with rough dollar equivalents. Paid plans are activated by our team, so upgrading starts with a message through the contact page. Full pricing and FAQ