--- title: "AI Lip Sync: Make Any Photo or Video Talk | Made" description: "Make any photo or video talk. Type the words or add your voice, then lip sync with Veo, Kling, OmniHuman, Sync, Happy Horse, WAN or VEED Fabric. Priced per second, shown before you generate." url: https://www.made.app/tools/lipsync --- Lip Sync Studio [UGC](https://www.made.app/ugc) ![Maya, creator](https://www.made.app/images/tools/lipsync/presets/maya.jpg) **Maya**Creator Change **Upload a photo or video** Drag and drop or click to browse What should they say?65/400 About 6s of speech Voice and delivery Optional Model Veo 3.1 Lite Length Resolution Veo 3.1 Lite · billed per second of video · failed videos are free Lip Sync Studio # Make any face say anything. Type the words or use your own voice, then choose from 16 models, Veo and Kling to OmniHuman and Sync. The price shows before you generate. 1 **Add a face**A photo or a short clip, or one of ours 2 > *Script* > > “Okay, real talk. I didn’t film this. It’s one photo and a script.” > > 6s · Veo 3.1 Lite **Give it words**Type a script, upload audio, or record 3 **Pick a model and generate**See the price first, then download the MP4 ## One photo, one line, nine models Maya says “Okay, real talk. I didn’t film any of this.” on every model that voices a typed line. Tap the speaker to hear each one. **Veo 3.1 Lite** 4 s **Veo 3 Fast** 4 s **WAN 2.5** 5 s · 480p **Happy Horse 1.1** 4 s · 720p **Kling 2.6** 5 s **Kling 3 Pro** 4 s **FLUX 3** 5 s **Veo 3** 4 s **Veo 3.1** 4 s ## One photo, one voice track, six models The same photo lip-syncing a 4.7-second voice track on every model that takes one. Tap the speaker to hear each one. **Kling Avatar 2.0** 4.7 s **VEED Fabric 1.0** 4.7 s · 480p **Creatify Aurora** 4.7 s · 480p **WAN 2.5** 5 s · 480p **Kling Avatar 2.0 Pro** 4.7 s **OmniHuman 1.5** 4.7 s · 1080p ## Made in Lip Sync Studio Real outputs, not demos. Recreate loads the face, the words, and the model, so you can run it yourself or swap in your own. **Maya** Creator “Okay, real talk. I didn’t film this. It’s one photo and a script.” Veo 3.1 Lite Script · 6 s **Lucía** Travel vlogger · Spanish “¡Hola a todos! Hoy les enseño mi café favorito aquí en la Ciudad de México.” Happy Horse 1.1 Script · 8 s **Vera** Oil painting “Five hundred years of sitting still. You have no idea how good it feels to talk.” Veo 3.1 Script · 8 s **June** Shop owner “This serum sold out twice last month. Here’s why people keep coming back for it.” VEED Fabric 1.0 Voice track **Theo** Podcast host “Welcome back to the show. Today we’re turning one photo into a whole episode intro.” Kling 3 Pro Script · 8 s **Biscuit** Dog “I did not eat the sandwich. The sandwich simply… left.” Veo 3.1 Lite Script · 6 s **Leo** Spokesperson “Give me thirty seconds, and I’ll show you exactly how this works.” Veo 3.1 Lite Script · 8 s **Leo** Spokesperson · Spanish “Hola, soy Leo. En los próximos treinta segundos, te enseño exactamente cómo funciona.” Kling LipSync Re-voiced clip **Leo** Spokesperson · Spanish Sync Lipsync 3 Re-voiced clip **Dev** Coach “Three minutes a day. No equipment, no excuses. Let’s get moving!” FLUX 3 Script · 6 s **Rosa** Storyteller “The secret to my sauce? Patience. And a little more garlic than you think.” Veo 3.1 Lite Script · 8 s **Nova** 3D character “Houston, everything up here is perfect. Also, I may have lost a sock.” Veo 3.1 Lite Script · 6 s **Your face next** A selfie, a product shot with a person, a pet, a painting, or a short clip you already have. ## Every model, side by side Each one bills per second at its own price, in credits. Unused credits roll over, and failed videos are free. | Model | Starts from | Voice | Length | Resolution | Good for | | --- | --- | --- | --- | --- | --- | | Veo 3.1 Lite Google | Photo | Writes it from your script | 4–8 s | 720p or 1080p | Quick, natural talking clips in 720p or 1080p | | Veo 3 Fast Google | Photo | Writes it from your script | 4–8 s | 720p | Veo 3’s voice, rendered faster | | WAN 2.5 Alibaba | Photo | Your script or a voice track | 5 or 10 s | 480p or 720p or 1080p | Speaks a line or your voice track, from 480p | | Happy Horse 1.1 Alibaba | Photo | Writes it from your script | 3–15 s | 720p or 1080p | Speaks the script in its own language, up to 15 seconds | | Kling 2.6 Kling AI | Photo | Writes it from your script (English and Chinese) | 5 or 10 s | 1080p | 1080p in 5 or 10 seconds. Speaks English and Chinese | | Kling 3 Pro Kling AI | Photo | Writes it from your script (English and Chinese) | 3–15 s | 1080p | 1080p for up to 15 seconds. Speaks English and Chinese | | FLUX 3 Black Forest Labs | Photo | Writes it from your script | 5–20 s | 720p | Long takes, up to 20 seconds | | Veo 3 Google | Photo | Writes it from your script | 4–8 s | 720p | The model that made talking video believable | | Veo 3.1 Google | Photo | Writes it from your script | 4–8 s | 720p | Google’s flagship, with the most natural delivery | | VEED Fabric 1.0 VEED | Photo | Your voice track, or a voice made from your script | As long as the audio, to 60 s | 480p or 720p | The photo speaks the voice track, as long as it runs | | Kling Avatar 2.0 Kling AI | Photo | Your voice track, or a voice made from your script | As long as the audio, to 60 s | 720p | A photo that speaks any voice track, up to a minute | | Kling Avatar 2.0 Pro Kling AI | Photo | Your voice track, or a voice made from your script | As long as the audio, to 60 s | 1080p | Kling’s talking avatar in 1080p | | OmniHuman 1.5 ByteDance | Photo | Your voice track, or a voice made from your script | As long as the audio, to 60 s (30 s at 1080p) | 720p or 1080p | Gesture and expression that follow the voice | | Creatify Aurora Creatify | Photo | Your voice track, or a voice made from your script | As long as the audio, to 60 s | 480p or 720p | A talking photo in 480p or 720p | | Kling LipSync Kling AI | Video | Your voice track, or a voice made from your script | Clips of 2–10 s | Your clip’s | Re-voices a clip you already have. Only the mouth changes | | Sync Lipsync 3 Sync | Video | Your voice track, or a voice made from your script | Clips of 2–60 s | Your clip’s | Re-times the mouth in clips up to 4K | [See every model’s price](https://www.made.app/pricing) ## Questions **What is AI lip sync?** It makes a face in a photo or video speak, with the mouth moving to the words. In Lip Sync Studio you either type the words or bring a voice track, then pick the model that animates the face. **Can I use my own voice?** Yes. Switch to Voice, then upload audio or record in the browser. MP3, WAV, M4A voice memos, and the sound of a video all work. Photo models like OmniHuman 1.5, Kling Avatar 2.0 and VEED Fabric make the photo speak your track; Kling LipSync and Sync Lipsync 3 re-voice a video you already have. **Do I need a voice recording?** No. Type the script and every model works. Veo 3.1 Lite, Veo 3 Fast, WAN 2.5, Happy Horse 1.1, Kling 2.6, Kling 3 Pro, FLUX 3, Veo 3, Veo 3.1 voice the words themselves; the voice-track models get a voice made from your script first, from 21 ElevenLabs voices you can preview. **Which model should I pick?** For a typed line, Veo 3.1 Lite is the least expensive, Veo 3.1 has the most natural delivery, and Happy Horse 1.1 speaks many languages for up to 15 seconds. To make a photo speak a voice track, Kling Avatar 2.0 costs the least per second and OmniHuman 1.5 adds the most gesture. To re-voice a clip, Kling LipSync bills in 5-second blocks and Sync Lipsync 3 handles clips up to 4K. The comparisons above run the same photo on each. **How much does lip sync cost?** Each model has its own price per second, in credits, and the exact price is on the Generate button before you start. A voice made from your script adds a small text-to-speech charge. Unused credits roll over, and failed videos are free. Every model’s price is on the price sheet. **What languages can it speak?** Write the script in the language you want spoken. Happy Horse 1.1 is built for multilingual speech, and the Spanish examples above were made with it and with a Spanish voice track. Kling speaks English and Chinese. With your own voice track any language works, because the mouth follows the audio. **How long can a video be?** Typed lines run 3 to 20 seconds depending on the model: Veo up to 8, Happy Horse and Kling 3 Pro up to 15, FLUX 3 up to 20. With a voice track, the photo models make a video as long as the audio, up to 60 seconds (30 seconds for OmniHuman at 1080p). Kling LipSync takes clips of 2 to 10 seconds and Sync Lipsync 3 up to 60. **Can I lip sync a video I already have?** Yes. Upload the clip and add the new voice track, or type the new line and pick a voice. Kling LipSync needs a 720p or 1080p clip of 2 to 10 seconds; Sync Lipsync 3 takes up to 60 seconds and 4K. Only the mouth changes, so you can fix a line or dub a clip into another language. **What can’t I make?** Only use faces and voices you have the right to use: your own, people who agreed, or characters you created. Don’t make someone appear to say something they didn’t say. The models decline some requests, for example ones involving minors or sexual content, and a declined video is never charged. ## More video tools [**AI video generator** Write a scene and generate it with sound.](https://www.made.app/tools/ai-video-generator) [**Video face swap** Put a new face on a clip you already have.](https://www.made.app/tools/video-face-swap) [**Motion transfer** Copy the moves from one video onto a photo.](https://www.made.app/tools/video-motion-transfer) [**UGC ads** Turn a product link into talking ads.](https://www.made.app/ugc) --- Markdown version of https://www.made.app/tools/lipsync. Every public Made page has one at its URL plus .md (https://www.made.app/tools/lipsync.md). Site summary for AI assistants: https://www.made.app/llms.txt