VEED Avatars: the mouth pass after the picture exists (4K, 30cr)
VEED Avatars wants a plate and a track. It is not how you invent the scene.
VEED Avatars wants a plate and a track. It is not how you invent the scene.
VEED Avatars is VEED's premade-avatars row: lipsync content type, audio on, 4K ceiling, 30 credits. The catalog description is "Professional talking avatar generation with natural expressions and lip synchronization." Features listed: Professional, Expressive, Natural, Avatar, High-quality. The plate here is a premade avatar, not a set you prompted from nothing. The track is the line that avatar has to speak. This is not text-to-video.
A premade face, not a new location
You do not use this row to invent a kitchen, a city, or a product orbit. You pick a talking avatar that already exists as the performer, then you give it audio (or the words that become audio). Lipsync is the job: the mouth has to match the track. AI lipsync and the AI avatar generator are the doors. A cinematic scene with no speaker is a different generate.
The catalog does not require an image. You are not bringing your own footage the way a video-to-lipsync row does. You are booking a premade performer. If the job is "re-voice this take we already shot," that is a video-to-lipsync model. If the job is "a professional talking avatar, 4K, 30 credits," this is the row.
4K, 30 credits, audio on
Thirty credits is the catalog price. Max output resolution is 4K. Audio is on because a talking avatar without a track is a still with a job title. There is no duration menu on the record; length follows the performance you ask for, not a 5s/10s picker like a text-to-video row.
Do not spend 30 credits to discover you actually wanted b-roll. Do not spend 30 credits to discover you wanted your own actor's plate. The VEED hub is the rest of that brand. This page is only Avatars: premade talking performer, lipsync, 4K, 30 credits.
Scene first is the wrong order — performer first is the right one
"The picture exists" on this row means the avatar exists. You do not generate a silent scene and then hope VEED Avatars will inhabit it. You pick the avatar, you pick the line, you run the mouth pass. If you also need a room around them, that is a different shot, a different model, a cut. Best talking-head models is the wider map.
Natural expressions and lip synchronization are the listed promise. They apply to the premade performer. They do not apply to a product you never uploaded, because this is not reference-to-video.
Not a scene generator
The test is whether the cut is a person talking. If it is not, leave. If it is, and you do not have your own footage, VEED Avatars is the 30-credit, 4K, premade-avatar answer. If you do have footage, do not force it through a premade avatar. Use a lipsync row that takes the plate.
FAQ
Can VEED Avatars generate an arbitrary scene from a prompt?
No. Category is premade-avatars. You get a professional talking avatar with lip sync, not a location you invented in text.
Is this the same as Kling-style video-to-lipsync?
No. Video-to-lipsync wants your existing footage. This row wants a premade avatar plus a track. Same mouth job, different plate.
Do I need to upload an image?
The catalog does not require an image. You are booking a premade performer, not driving your own still, unless the tool flow asks for a face on a different avatar row.
What are the 30 credits for?
The catalog price for this talking-avatar generate. 4K ceiling, native audio on the record, not a cheap text-to-video test.