Re-voice AI Video with sync-lipsync v2 on Flovaly

Re-voice is the Studio mode that gives an existing video a new voice: upload a clip, write a new script, pick a voice, and Flovaly re-synchronises the speaker's lips to the new audio while leaving the rest of the footage untouched. It is powered by sync-lipsync v2 Pro from sync.so and costs a flat 25 credits per clip.

What is sync-lipsync v2?

sync-lipsync v2 Pro (also written lipsync-2 Pro) is sync.so's audio-driven video model. Given a video of a person speaking and a replacement audio track, it regenerates the mouth region so the lips match the new words — frame by frame — while preserving the speaker's appearance, the camera work, and everything else in the shot. It is positioned as one of the most accurate re-sync models available, and it is what Flovaly runs for every Re-voice job.

The distinction from Talking Actors is the starting point. Talking Actors generates a brand-new video from a still image and audio (that pipeline runs on OmniHuman 1.5 — see our Talking Actors explainer). Re-voice keeps your existing footage and changes only what is said.

What re-voicing is good for

• Localisation. One clip, many markets: re-voice the same footage into another language or accent without re-shooting or re-generating. Any voice in the picker works, and the script drives the language — see the full localisation workflow for the step-by-step version. Flovaly's curated voice list is English-accented only rather than a dedicated multi-language library — for heavy localisation at volume, see how Flovaly compares to a platform built for that in Flovaly vs HeyGen.
• Fixing a take. The visuals are right but a line is wrong — a mispronounced product name, an outdated price, a claim legal wants changed. Rewrite the script and keep the footage.
• A/B-testing hooks. The first three seconds decide whether an ad gets watched. Re-voice lets you run five different opening lines over the same clip and let the numbers pick the winner — far cheaper than generating five full videos, and far cheaper than re-booking a real creator for each variation (see AI actors vs real UGC creators for the fuller cost comparison).
• Repurposing footage. A clip that worked for one product or offer can carry a new message without losing the performance that made it work.

How to use Re-voice on Flovaly

1. Open the Studio and switch to the Re-voice tab in the composer mode row.
2. Drop in a video, or click to upload — MP4 or MOV, up to 100MB. Short clips are the sweet spot; most re-voice jobs are under 30 seconds. The cleaner the source — steady framing, a clearly visible speaking face — the cleaner the re-sync.
3. Write the new script. The limit is 1,000 characters, with a live counter — write it the way people speak, since punctuation drives the pacing of the generated voice.
4. Pick a voice from the curated list; each entry shows its name, accent, and style so you can match the on-screen speaker.
5. Decide on captions. The Burn captions into the output toggle is on by default — captioned clips perform better on short-form social — and one click turns it off.
6. Generate. The job runs in the background and the re-voiced clip lands in your gallery when it finishes, ready to download or share.

Constraints and pricing

Re-voice is priced at a flat 25 credits per clip — no duration or resolution maths, because the output matches your source video. The practical limits: MP4 or MOV input up to 100MB, a 1,000-character script, and a clearly visible speaking face in the source footage (the model needs a mouth to re-sync). As with every generation on Flovaly, failed jobs are refunded automatically. For how credits map to plans, see Credits, plans, and billing.

Re-voice vs generating fresh

A fair question: at 25 credits, when is re-voicing better than generating a new clip? The answer is whenever the footage itself is the asset. A take with a great performance, a clip that has already proven itself in an ad account, real filmed footage you cannot regenerate — those are re-voice jobs. If you are still hunting for the right visual, generate fresh instead: Seedance 2.0 Pro starts at 30 credits and gives you a new shot each run — see our Seedance 2.0 Pro explainer.

When to choose Re-voice

Choose Re-voice when you already have the right video and the wrong words: localising a winner into new markets, fixing a single line, or iterating hooks over proven footage. Choose Talking Actors when you are starting from a character rather than a clip, and a base video model when the visual itself needs generating — our 2026 AI video model comparison covers those options side by side. And whichever pipeline you use, the same delivery habits apply — see AI UGC ads that don't look AI for what makes a re-voiced line sound spoken rather than read.

FAQ