Video Localization with AI

What's automated — and what isn't

Automated

Transcription, translation into 21 languages, AI voice track, subtitle files, timestamped transcripts

Still manual

On-screen text and graphics, cultural references, units and currencies, final native-speaker review

How teams use it

• Marketing: test explainers and product videos in new markets.
• Training: give every site the same onboarding and safety content.
• Courses: sell existing courses to new audiences.
• Webinars: turn live events into on-demand content per region.

A simple localisation workflow

1. Pick evergreen videos and the markets that matter.
2. Translate one video per language and have a native speaker review it.
3. Fix your glossary, then process the rest.
4. Translate titles, descriptions and thumbnails, and publish with subtitles.

The full list is in the video localization checklist.

Pricing

Transcript and subtitles

1 credit per minute (0.1¢), minimum 5 credits per file

Translated voice (includes everything above)

50 credits per minute (5¢), minimum 2 minutes per file

Credits

$1 buys 1,000 credits; credits don't expire

Free credits

100 credits when you sign in with Google

Privacy

Your video is never uploaded — only its audio is, over HTTPS in small chunks. Audio, transcripts and voice files are deleted after delivery, within 30 minutes at most, and are never used to train models. Your account keeps only the file name, length, languages and credits used. Details in the privacy policy.

Good to know

• One of 8 stock voices reads the whole translation — no voice cloning and no lip-sync.
• Music and sound effects are kept by AI vocal separation, which isn't perfect: in dense mixes faint traces of the original voice can remain, and sung lyrics are removed along with the speech.
• No speaker labels — transcripts don't say who is speaking.
• Files only: no links from YouTube or social apps, and no live audio.
• Subtitles come as SRT/VTT files rather than burned into the picture.

Who runs localization with a tool like this

The typical user is a small team without a localization vendor: a marketing manager taking product videos into two new markets, an L&D lead with onboarding content for sites abroad, a course creator testing demand in another language. They need the spoken layer handled and can find one fluent reviewer per market. Larger teams use the same output to draft and test a market before committing studio budget.

What it doesn't replace is the rest of a vendor's service: project management across many files, translation memory that reuses approved wording, casting voice actors, or adapting graphics. If your project depends on those, use this for the speech layer and plan the rest separately.

Tiering a video library

Not every video deserves the same treatment. Sorting the library first keeps cost and review effort where viewers will notice.

Tier 1: flagship brand films and paid ads

AI version to test the market; professional review or voice-over before full spend

Tier 2: product explainers, onboarding, training

Dub mode plus a fluent reviewer's pass

Tier 3: webinars, event recordings, internal updates

Translated subtitles from transcript mode; add a dub only if viewers ask for it

Tier 4: archive and low-traffic videos

A transcript for search, or leave as is

Building a localization strategy covers auditing a library and choosing markets.

Budgeting per market

The same library as translated subtitles only is 126 credits (12.6¢). At these rates the processing cost is rarely what limits a project; reviewer time is. Plan roughly one reviewer session per language per batch, and process the rest only after the first video's feedback is in. Running one video through several languages covers file naming and release order.

What sits outside the speech layer

• Text in the picture: lower thirds, slides and interface labels stay in the original language. Exporting textless versions of your videos makes replacing them far easier; see textless masters.
• Metadata: titles, descriptions, thumbnails and landing pages aren't translated by mydubly.
• Prices, units and dates spoken in narration: machine translation renders them rather than adapting them, so a price in dollars stays in dollars. Reviewers should flag every one.
• Product demos: the narration changes language while the interface on screen doesn't; localizing product demos covers how to handle UI terms.
• Cultural fit: humour, examples and references may need rewriting at the source, not fixing in translation.

Review and upkeep across languages

1. Write a one-page style guide per language: formality, terms to leave untranslated, how to handle names. Writing a translation style guide has a template.
2. Send each reviewer the SRT and the video. They note problems with timestamps, which makes fixes quick to locate.
3. Fix subtitle text directly in the SRT. For a mistranslated line in the dub, replace that line in an editor using the translated audio, or decide the error is minor enough to accept.
4. When the source video changes, re-run it for every language that carries it, and keep a list of which version each market has; updating translated videos covers patching versus re-running.

Deciding whether a market is worth continuing

Treat the first batch in a new language as a pilot with a fixed review date. Publish a handful of Tier 2 videos, then compare them with the same videos in the original language: watch time, how far viewers get, comments in the new language, sign-ups or support tickets from that region. Subtitles are the cheaper first test; if viewers in that market stay longer and ask for audio, add the dub later from the same source files.

Set the decision rule before you look at the numbers, such as "expand if the translated versions hold viewers about as well as the originals", so a slow first month doesn't end a market that needed time. Measuring localization success lists metrics by goal.

Where to go from here

Use-case pages show the workflow for specific content: marketing videos, training and online courses. The target-language pages, for example German or Portuguese, list what reviewers in that language should check.