Dubbing a Video into Another Language

The Dub panel turns the speech in one of your clips into another language. It transcribes what’s said, translates the script, re-voices each line with an AI voice at the moment it was spoken, and adds translated captions. For a video of someone talking, it can also re-render the speaker’s mouth so their lips match the new words.

Dubbing is on every plan and uses AI credits for each step. See What it costs.

What you can dub

Video clips you uploaded or recorded
Audio clips with speech, such as recordings and voiceovers, whether they’re on a scene or attached to an item

Dubbing works on your own media only. Clips from the stock libraries can’t be dubbed.

You can dub into 20 languages: English, Spanish, French, German, Italian, Portuguese, Dutch, Russian, Arabic, Hindi, Chinese (Simplified), Japanese, Korean, Turkish, Polish, Swedish, Indonesian, Vietnamese, Thai and Norwegian.

Open the Dub panel

The Dub panel works on one clip at a time, so you open it from that clip:

A video clip: select it on the canvas and click Dub in its toolbar
An audio clip on a scene: click the clip on the timeline, then click Dub in the toolbar above the canvas
An audio clip attached to an item: select the item, click the clip in its audio row on the timeline, then click Dub at the top of the clip’s settings panel

You can also open the AI Video panel, choose Dubbing, and click Open the Dub panel, or press Ctrl+P (Cmd+P on a Mac) and choose Dub Video Into Another Language. Both use the selected video clip.

The panel stays on the clip you opened it from, even if you click other things while it works. The clip’s name is shown at the top of the panel.

Dub a clip

Under Listen to the original, click Transcribe video (or Transcribe audio). Glintia transcribes the speech with its exact timings and splits it into spoken lines
Choose a Dub language. Spanish is selected to start with
Choose a Dub voice. Each voice is listed with its name and gender. One multilingual voice speaks every language on the list, so the speaker sounds like the same person in any dub you make
Optionally, type a direction in the box below, such as “keep it formal” or “keep product names in English”
Click Translate into your language, for example Translate into Spanish
Review the script. Each line shows its timing, the original words, and the translation, which you can edit
Set Keep original audio under the dub, from 0% to 60%. Leave it at 0 for a clean dub, or raise it a little to keep the room tone and music
Check the credit estimate next to Voice, then click Generate the Spanish voice (the button names your language)
For a video, click Match the mouth to the dub if you want the speaker’s lips to match. This step is optional

While the voice is generated, the panel shows its progress (“Voicing 3 of 20 lines…”). When it’s done, the clip plays the dub.

What you get

The dubbed voice: each line becomes a voice clip, placed at the moment the original line was spoken. For a video or an item’s clip, they’re attached to that clip’s item. For a scene clip, they’re added to the scene
The original, turned down: the clip’s own audio plays at the Keep original audio under the dub level. It isn’t deleted, so you can change the balance at any time with the same slider
Translated captions: a subtitle track named after the language, such as “Dub captions (Spanish)”, shown with a caption design over the part of the clip that has speech. To edit the lines or pick another design, open the Subtitle panel and choose that track

Switch between the dub and the original

Once a clip has a dub, a switch with Original and the dub’s name (for example Spanish dub) appears in the clip’s toolbar (for an item’s audio clip, in its settings panel) and at the top of the Dub panel. Click it to flip between the two. With the original playing, the dubbed voice is silent and the translated captions are hidden. Switching is instant and costs nothing, and your export uses whichever version is selected.

The Dub button in a dubbed video’s toolbar shows a dot, and its tooltip says which language the clip has a dub in.

Fix lines and generate again

You don’t have to start over to fix a line:

Edit a translation: type in the line’s box. The line loses its old voice, and the button changes to Generate the changed lines. Clicking it re-voices only the lines you changed, so you only pay for those
TOO LONG: some translations take longer to say than the original. A line marked TOO LONG runs into the next one, and the panel tells you how many lines to shorten. Shorten them and generate again
FAILED: if a line couldn’t be voiced, click Retry … failed lines
VOICED: the line has a current voice

If some lines didn’t translate, the panel says how many were translated. Fill in the rest by hand, then generate the voice.

To dub into a different language, choose it under Dub language and click Re-translate into that language. A clip holds one dub at a time, so this replaces the current script, and generating again replaces the dubbed voice. If you change the language but don’t re-translate, the panel reminds you that the script is still in the old language.

Match the mouth

For a video clip, Match the mouth re-renders the speaker so their lips say the dubbed words.

It’s available once at least one line has a voice
It takes a few minutes. You can keep working while it runs
When it’s done, the panel says “The mouth now matches the dub”, and the clip plays the re-rendered video with the dubbed voice built in
If you change a line and generate it again, the clip goes back to the original footage under the new voice, and you need to run Match the mouth to the dub again

If the panel shows “Mouth matching timed out. The job may still finish — reopen this panel later.”, give it some time and open the Dub panel on the clip again.

What it costs

Each step uses AI credits from your monthly allowance:

Step	AI credits
Transcribe	0.25 per minute of audio, minimum 0.25
Translate	1 per translation of the whole script
Generate the voice	0.5 per second of generated speech
Match the mouth	3 per second of video

The panel shows an estimate before you generate the voice, and the charge is based on the real length of the speech that comes back. The Lip-sync line shows the most that matching the mouth can cost (“up to … credits”). If you run out of credits, see Plans, Limits and Usage.

Your work is saved on the clip

The transcript, translation, voices and settings are saved with the clip. If you close the panel and open Dub on the same clip later, you pick up where you left off.

Troubleshooting

“Only your own uploads and recordings can be dubbed — this clip comes from a stock library.”: upload your own video or recording and dub that
“No speech was found in this clip.”: the clip has no clear speech for the transcription to pick up
Transcription fails on a large file: transcription accepts source files up to 25 MB. Compress the file before uploading it, then try again
“You have reached your AI credit limit for this month”: wait for your allowance to renew, or see Plans, Limits and Usage
The script is in the wrong language: you changed Dub language after translating. Click Re-translate into the new language before generating the voice