# Lip Sync Source: [Lip Sync]() Documentation index: [llms.txt]() Last updated: 2026-11-05 Lip Sync changes the mouth of the person in your own video so it matches speech audio you supply. The picture, the person, and the scene stay yours. Only the speaking changes. Use it when you have replaced a line of dialogue, re-recorded a take, or generated new speech, and the old mouth movement would no longer fit. ## When should I use Lip Sync? Use it when the footage stays and the words change, and you want the mouth to follow the new words. Pick another feature in these cases: - **You want the same video in another language.** Use [AI Video Translation](https://pr-798-preview.chatcut.dev/docs/video-translation). It translates the speech, and mouth synchronization is one of its options, so you do not run Lip Sync on top of it. - **You have no footage, only a photo or a script.** Use [AI Avatars / Digital Humans](https://pr-798-preview.chatcut.dev/docs/digital-humans) to create a new presenter performance. - **You only need new narration and the picture can stay as it is.** Use [AI Voiceover](https://pr-798-preview.chatcut.dev/docs/ai-voiceover) and replace the audio. The mouth will not match the new words, so this suits footage where no speaker is visible or the mismatch does not matter. - **You want a different voice without touching the picture.** That is a voice job, not a lip sync job. Lip Sync does not create or clone a voice. Lip Sync does not translate, write dialogue, or generate speech. It only moves the mouth to audio that is already final. ## What do I need before I start? - **A video with one clearly visible speaker.** The face should be in view and not hidden. The video must still have its original speech audio, because Lip Sync uses it to find who is speaking and when. - **The final speech audio.** This is the exact audio you want heard in the result, already in your project. If you do not have it yet, ask the Agent to generate it with [AI Voiceover](https://pr-798-preview.chatcut.dev/docs/ai-voiceover) first, or import a recording. - **Matching lengths.** The audio and the video need to be close in length. See the limits below. - **Only the part you want changed.** Lip Sync processes the whole asset you give it, so export that section as its own asset first. - **Credits and a plan that includes it.** See [Credits](https://pr-798-preview.chatcut.dev/docs/credits-policy) and [Free and Pro](https://pr-798-preview.chatcut.dev/docs/free-and-pro). ## How do I ask the Agent? 1. Import the video and the speech audio into **My Assets**, or make sure both are already in the project. 2. Switch the AI composer to **Agent** mode. 3. `@` mention the video and the audio, and say that you want the mouth to match the new speech. 4. Say whether the video may stretch a little to fit the audio, or must keep its exact length. 5. Send the request. The Agent checks both lengths and that the video has audio, then shows a confirmation card. 6. Play the audio on the card. When the words and voice are right, select **Confirm and generate**. Select **Not now** to stop. 7. Follow progress in the conversation and **My Assets**. The result arrives as a new video asset. Drag it to the timeline when you are ready. ## Prompts you can copy Prompt ```text Make the speaker in @interview-take-3 say the audio in @corrected-line instead of the original line. Sync the mouth to the new audio. Keep the video length exactly the same, and save the result as a new asset without replacing the original. ``` Prompt ```text Use @product-demo-presenter and @new-voiceover. Sync the presenter's lips to the new narration. It is fine if the video stretches slightly to fit the audio. Put the result on the timeline after the current clip. ``` Prompt ```text Generate a voiceover that says: [paste the final line]. Let me listen to it first. After I approve it, lip sync @host-clip to that voiceover. ``` ## What does ChatCut do for me? - **Checks the inputs.** It confirms that the first input is a video that has audio, that the second is audio, and that the lengths fit. - **Shows you the speech first.** The confirmation card plays the actual audio so you can check the words and the voice before anything is generated. - **Shows the cost first.** The card carries the credit estimate. Lip Sync always asks each time. Auto-Allow does not skip it. See [Generation Confirmations](https://pr-798-preview.chatcut.dev/docs/generation-confirmations). - **Delivers a new asset.** Your original video and audio stay untouched. ## What do I get? You get a new video of the same scene in which the speaker’s mouth follows the audio you approved. It is named after the source video. Lip Sync re-renders the picture, so the file is not guaranteed to be identical to the original in codec, frame rate, or transparency. If the picture must stay exactly as shot, replace the audio only. ## What are the limits and tips? - **One visible speaker.** Lip Sync has no way to pick one person from several. Cutting a time range does not hide the other people still in frame. Export a shot with one speaker as its own asset first. - **Lengths must be close.** With an exact length, the audio and video must be within a quarter of a second. If the video may flex, the longer one can be at most about 15% longer than the other. Outside that range, the Agent asks you to trim or adjust the audio first. It will not cut your approved words. - **Fix the audio before the mouth.** Wrong words, pronunciation, or voice come from the audio. Lip Sync cannot repair them, so correct the audio and run it again. - **Check the footage if the mouth looks off.** Hidden faces, fast scene cuts, and several people talking make the result less natural. Try a cleaner shot. - **Review with sound on.** Play the result in full and watch the mouth, not a still frame. - **Equal length is not alignment.** A pause or a breath in the wrong place still looks off. Match the timing of the sentences, not only the total length. - **A video is processed whole, and billed on length.** The estimate depends on the video and audio length, so use short clips while you test. - **Up to 10 video generations at once.** This count is shared with video generation and video translation. Wait for a running job before sending more. ## Where is Lip Sync available? Lip Sync is in the web app and ChatCut Desktop, and agents connected to ChatCut can request it too. It is a Pro feature. If your plan does not include it, ChatCut says so before generating and shows how to upgrade. Your credits are kept. ## What related pages can I read? - [AI Video Translation](https://pr-798-preview.chatcut.dev/docs/video-translation) - [AI Avatars / Digital Humans](https://pr-798-preview.chatcut.dev/docs/digital-humans) - [AI Voiceover](https://pr-798-preview.chatcut.dev/docs/ai-voiceover) - [Generation Confirmations](https://pr-798-preview.chatcut.dev/docs/generation-confirmations) - [Credits](https://pr-798-preview.chatcut.dev/docs/credits-policy)