Talking avatar · fal
OmniHuman 1.5
A photo and a voice track become a person speaking.
15 creditsper 5 s · from $1.60 · 3 cr/second
What it does
One photo plus a voice track. Output length follows the audio: up to 30s at 1080p, 60s at 720p. Emotion and gesture track the voice, not just the lips.
What you give it
- Takes
- Image, Audio, Text prompt (optional)
- Produces
- video
- Length
- 1–60s
- Reference images
- 1
- Resolutions
- 720p, 1080p
- Audio
- keeps the source track
- Lip-sync
- yes
- Category
- Talking avatar
- Runs on
- fal · fal-ai/bytedance/omnihuman/v1.5
Required: image, audio. Anything else listed above is optional.