Opening ai lip sync.
0.0s
0.0s
Video studio · Lip sync
Upload a clip you already filmed. Attach a voice track, or type the line instead. The engine redraws the mouth until it agrees with the sound. Same take, same framing, new words.
Good evening. Tonight's story starts, as they all do, with a door left open.
Every word you hear was matched to this face, frame by frame.
No audio yet? Write the script in Text to Speech, cast a voice, and sync it here.
01 · The bench
Every engine reads a mouth with its own hand. A clip that lands oddly on one chair often locks cleanly on the next, so the studio keeps your clip and your track loaded while you change seats. A second opinion costs one click, not a re-upload.
Kling Lipsync Text takes a typed line instead of an audio file and speaks it straight into the clip, keyboard to take. Nothing was ever recorded.
02 · Field work
Every version keeps the take exactly as filmed. The only thing that changes between markets is the voice lane.
Keep the performance, change the language. The new track carries the words and the sync pass makes the face agree with them. What the camera caught stays caught.
Hand a generated presenter the exact script. The avatar reads it take after take, without a camera in the room.
Film the walkthrough once, then publish a version for every market it reaches. One shoot, a shelf of takes.
03 · The supply line
The voice lane takes three feeds. A track you already recorded. A line you type straight into the Kling text seat. Or a narration you build one door down: write the script in Text to Speech, pick the voice, and carry the finished track here. The two tools were built to run as one line.
It takes a video you already have and a new audio track, then redraws the mouth so the face speaks those exact words. The rest of the clip stays as filmed: same take, same framing, new line.
That is the whole job. Upload the clip, attach a voice track, and the engine syncs the audio to the video by matching the mouth to the sound. Several engines are on the bench, so if one read feels off you can run the same clip through another.
No. Kling Lipsync (Text) takes typed text instead of an audio file, so the clip speaks a line that was never recorded. And if you want a specific narration, write the script in the Text to Speech studio, pick a voice, and bring the track here.
Sync Lipsync 2 is the everyday first chair, with Sync Lipsync 3 and Sync Lipsync 2 Pro in the same family. Veed, Kling, LTX Video 2, Pixverse and ByteDance fill out the bench. The same clip can land differently on each, so a second take is often the fastest fix.
You get free tokens at signup, no card required, and lip sync jobs run on those like everything else in the studio. When you outgrow the free tokens, plans are one time payments, not subscriptions.
Yes. Keep the clip, swap the voice track, and generate a version for each market. The face and the performance stay; only the words change.
05 · Nearby
06 · Begin
Record it, type it, or generate it. The face will say it, on the take you already have.
Free tokens at signup · no card required · everything you make is yours