VEED Lipsync: the mouth pass after the picture exists (4cr)
VEED Lipsync wants a plate and a track. It is not how you invent the scene.
VEED Lipsync wants a plate and a track. It is not how you invent the scene.
VEED Lipsync is fast, affordable video lipsync: simple video plus audio in. Category is video-to-lipsync. Content type is lipsync. Audio is on. An image is not required — because the plate is a clip, not a still. The catalog lists 4 credits. Features named: budget-friendly, fast, simple. No duration ladder, no resolution token, no aspect-ratio list on the slug. The matrix is per-second with a 4-credit floor and an 81-credit ceiling.
That is a re-speak pass. The scene already moved. You are changing what the mouth is doing, not where the camera is.
Video in, audio in
Simple video + audio input is the whole description. You already have a take: a talking clip whose line is wrong, a performance you want to dub, a UGC shot whose original VO you cannot ship. You already have a track: a TTS file, a new read, a localisation. VEED Lipsync combines them. It will not generate the café, the walk, or the product pickup.
requires_image is false and requires_video_input is not the story the categories tell: this row is video-to-lipsync. If you only have a still, you wanted image-to-lipsync (Wan 2.2 Speech to Video is that shape at 720p / 50 credits). If you only have a script and no picture, you wanted a premade avatar (HeyGen Avatar V3 at 17 credits) or a talking generate. VEED Lipsync is the cheap pass when both files already exist.
Four credits is the listed floor. Duration-driven billing means a long dub is a long bill, up to 81 on the matrix. Cut the audio to the line. Cut the video to the take. Then spend 4 credits on the mouth, not on three minutes of room tone.
The AI lipsync tool is the launcher. VEED's roster and best VEED model sit next to this slug. Best lipsync is the job map. Lipsync a video is the editing-task twin. This page is the 4-credit re-speak rule.
Fast is not a scene generator
Budget-friendly, fast, simple — those features are about a known take, not about inventing one. If you do not like the background, replace the clip first. If you do not like the wardrobe, reshoot or edit the picture. Then lipsync. Asking VEED Lipsync to "make it more premium" is a prompt aimed at a model that is waiting for two files.
No published resolution. Do not brief 4K against a blank field. Deliver what the row returns. If the master is 4K talking, this is the draft dub or a different model. No published aspect list: bring a clip already in 9:16 if the platform is 9:16. Lipsync will not be your crop tool.
Audio is the new line. It is not a music bed. If you also need a score, add it after the mouth is right. If you needed a silent plate, you are in the wrong building.
Re-speak is not a first take
A first take from text. A still that starts talking (wrong category). A 700-avatar cast (that is premade, not video-to-lipsync). A 50-credit 720p performance from a single photo (that is Wan).
VEED Lipsync is the row you run when the picture exists as motion and the words need to change. Skipping the picture is how 4 credits feels like a talking generate and then fails like a missing file.
Captions after. A dubbed mouth is not a subtitle track. Burn the new line once you would ship the take.
Clip plus track, or nothing
Do you have a clip you would keep if the line were right? Do you have a track that is the right line?
If both yes, this is the 4-credit mouth pass. Do not first ask it to change the room.
If you have only a still, change rows. If you have only a script, cast a twin or generate a talking shot. If you have neither, you do not have a lipsync job yet.
FAQ
Can VEED Lipsync start from a still?
Not as the described job. The category is video-to-lipsync: video plus audio. A still-plus-audio talking generate is image-to-lipsync on another slug. Bring a clip.
Why is this 4 credits when other lipsync lists 17 or 50?
4 credits is the listed floor for this simple video+audio row. HeyGen Avatar V3 is a premade-twin job at 17. Wan 2.2 Speech to Video is an image-required 720p performance at 50. Same mouth-pass idea, different plate. Pick the plate you actually have.
Does length have a picker?
The catalog does not list durations. Billing is per second with a 4-credit floor. Cut both files to the line you will ship. A long unmatched read is how "fast and affordable" becomes the 81-credit ceiling.
Will it invent a soundtrack besides the speech?
Treat the audio input as the line. This is lipsync, not a scorer. If you need music, add it after. If you have no video, you have nothing for the mouth to ride on — generate or shoot the plate first.