Photo + audio → talking video
OmniHuman 2
Upload one photo and one audio track, and get a video of that person speaking or singing.
Two inputs. No timeline, no keyframes, no separate singing mode.
One photo + one audio track
The photo sets the performer. The audio drives the performance.
How OmniHuman 2 works
Bring a face and a sound. The model performs the rest.
- 01
Upload one photo
A portrait, a selfie or any clear picture of one person, taken from the front with the face fully visible.
- 02
Add one audio track
A spoken line or a song. The audio is the performance, so what is said and how it is said both end up on screen.
- 03
Generate the video
Pick the clip length, choose the framing, and download the finished video when it is ready.
A plain background and a single clear face give the best results. Keep the audio clean and free of background noise.
Examples
Eight ways one photo and one audio track become a performance.
Hover a clip for its input recipe.
Sample clips on this page are AI-generated illustrations of the photo-plus-audio workflow. They were not produced from your uploads and are not a promise of a specific result.
What you get
Four things that decide whether this fits your workflow.
Photo and audio are the whole input
There is no prompt to write and no timeline to cut. Upload a picture of one person, add the track they should perform, and generate.
Speech and singing, one workflow
A spoken paragraph and a sung verse use the same two inputs. There is no separate singing mode to switch on.
A studio you can use today
The generator on this site is live. Open the studio, upload your photo and audio, and generate a clip instead of joining a waitlist.
Ten languages, end to end
The whole workspace, from the landing page to the studio and the support address, is available in ten languages.
Use cases
Anything with a voice and a face you already have.
Explainers and lessons
Turn an instructor photo and a recorded explanation into a short lesson clip without booking a shoot.
Songs and lyric clips
Give a singer's portrait a sung line for a lyric teaser or a music drop.
Podcasts and audiobooks
Add a face to a narrated passage so an audio-first project has something to show.
Social shorts
Produce vertical talking clips for feeds and stories from a photo you already own.
Product and ad lines
Read a single advertising line with a presenter portrait, then cut it into a longer edit.
Characters and mascots
Bring a drawn or sculpted character to the camera for a short line or a reaction shot.
Pricing
Every generation is charged in credits, and the cost is shown before you generate.
See pricingPlans and credit packs are listed on the pricing page.
Questions, answered
The things people ask before their first clip.
One picture of one person and one audio file. The picture establishes who appears and how the shot is framed; the audio is what they perform. Both are required.
Give your audio a face
One photo, one audio track, and a finished talking or singing clip.
Open the studio