AI Lip Sync
Make any photo talk
Upload a photo, type what it should say or add your own audio, and FunGen turns it into a talking video with natural lip sync, blinking and head movement.
Only the words to say, not instructions like "make him say…". Typed lyrics are read, not sung; to make the photo sing, upload a song with vocals.
The photo lip-syncs to the voice in your recording or song, so a song needs vocals (an instrumental won't make it sing).
A hint for how the person moves. It isn't followed exactly.
If the 1080p model is unavailable, we use a 720p backup and charge the 720p rate.
Three steps
Upload a photo
Any clear portrait: a selfie, a product mascot, an illustration or a 3D character.
Add the voice
Type the line and pick a voice, or upload your own recording up to 60 seconds.
Generate
FunGen syncs the lips to every word and adds natural blinking and head movement.
Who uses it for what
Want the full walkthrough?
How to make a photo talk with AI: photo tips, writing lines that sync well, and fixing common problems.
Questions about AI lip sync
What is AI lip sync?
AI lip sync animates a still photo so the mouth matches a voice track. FunGen also adds blinking, small head turns and expressions, so the result looks like the person is really speaking.
Do I need my own audio?
No. Type the line and pick a voice, and FunGen generates the speech for you. You can also upload an MP3, WAV or M4A recording, up to 60 seconds.
What photo works best?
A clear, front-facing photo with the face and mouth visible, waist up or closer. Illustrations, 3D characters and pets also work.
How many credits does it cost?
11 credits per second of video at 720p and 21 at 1080p, so a 10-second 720p clip is 110 credits. You see the estimate before you generate.
How long does it take?
Usually 3–5 minutes for a 10-second clip. Longer audio takes a little longer.
Can I lip sync a video instead of a photo?
Not yet on this page. For now it works from a single image; video-to-video lip sync is coming next.
Only animate photos you have the rights to. Don't make real people say things they didn't say without their consent.