VEED Lipsync is 4 credits on a video

VEED Lipsync is 4 credits on a video you already have. Fabric is 40 credits to animate a still. You are replacing the mouth. You are not building the talking take.

VEED Lipsync is video-to-lipsync. Audio on. Image not required. Sticker 4 credits. Description: fast and affordable video lipsync generation; simple video plus audio input. The talking take exists. The new line exists. This row maps the second onto the first.

VEED Fabric 1.0 is the image-to cousin in the same brand: still plus audio, 720p, 40 credits. Ten times the sticker because the picture does not exist as motion yet. Do not run Fabric on a clip. Do not run Lipsync on a JPEG.

The job test is one sentence: do you already have the talking take? Yes: 4 credits, this slug. No: you are in image-to-lipsync or a talking generate, and 4 credits will not invent the shoulders.

You already have the talking take

Video-to-lipsync is a redub. The body, the room, the hands, the cut — those pixels stay. The mouth follows a new track. Translated VO. A cleaner read of the same script. A different language on the same presenter. A line that legal changed after the shoot.

If any of those are the job, opening Fabric is the expensive wrong door. Fabric will try to create a talking clip from a still. It will not keep the camera move you already paid for. It will not keep the product in the hand. It will not keep the cut you already made. Forty credits to lose the take is not a savings versus 4.

Bring two files: the video, and the audio that should own the mouth. Play them against each other before you confirm. If the new track is longer than the picture, you will get a mouth that runs out of face or a clip that pads. Cut the line to the picture, or extend the picture on an extend row, then lipsync. This slug is not an extend.

The AI lipsync tool will list both VEED cards. Name video-to-lipsync. The lipsync video task is the editor phrasing. AI dubbing is the language job that often sits in front of this pass: make the track, then replace the mouth.

A still of the presenter is not “the talking take.” A still is Fabric or Kling Avatar Pro. A paragraph with no WAV is TTS. A silent generate you hoped would talk is a native-audio row, not a 4-credit redub.

Four credits versus forty is the category, not a discount

People treat 4 versus 40 as a VEED sale. It is not a sale. It is two categories.

You have	VEED row	Sticker	Ceiling
Clip + new track	Lipsync	4 credits	none listed
Still + track	Fabric 1.0	40 credits	720p

The cheap number is cheap because the model is not inventing motion, lighting, or a body. It is moving a mouth on pixels that already decided those things. If those pixels are wrong — bad light, wrong product, the hand in front of the lips — 4 credits will faithfully lipsync a bad take. Fix the take, or generate a new one, then come back.

Kling’s video-to cousin is Kling Lipsync: video-to-lipsync, 4K, 7 credits. Same category, a different ceiling and sticker. This page is the VEED 4-credit card. If you needed 4K on an existing clip, change the row. If you needed 4K from a still, that is Avatar Pro at 58, not this slug.

No duration chips on VEED Lipsync. Length follows the inputs. No max resolution listed. Output follows the plate. A 720p talking take comes back as a 720p talking take with a new mouth. Uploading a compressed story export and expecting a 4K master is a different job: generate or upscale, not this pass.

audio is true because the driving track is the point. The row does not write a score. Keep the mix in the editor if the source had room tone you still want. Replacing the mouth does not automatically replace the bed.

Replacing the mouth is a language and legal pass

The usual reason this row exists in a week is not “I want a talking AI.” It is “the picture is approved and the words are not.”

• Translation. The presenter stays. The line changes language. Make the VO on a TTS or clone row, then 4 credits on the mouth. Caption the new language after. Add captions to video.
• Legal. One sentence in a 20-second take is now forbidden. Re-record the sentence. Lipsync that region — or the whole take if the row does not let you mask a range. Do not regenerate the kitchen.
• Performance. The take is good; the read is flat. A new track on the same face is cheaper than a new shoot and cheaper than Fabric.

Photoreal mouths on real people are a disclosure problem when the audience would take the performance as authentic. EU deployers: visible or audible label at first exposure under Article 50 from 2 August 2026. New York ads with synthetic performers are a different statute. This row does not exempt you because it only moved the lips.

Health, finance, law, politics: YouTube’s inauthentic-content bucket for synthetic personas presenting as human authorities is the format risk, not the slug. A 4-credit redub of a real person reading a real script is still a real person. A 4-credit redub of a generated face reading medical advice is the bucket. Know which file you uploaded.

Silent autoplay still applies. The new mouth is useless if the first three seconds play without sound and without type. Captions for silent autoplay. Burned-in captions on a redub should match the new language, not the old one.

Do not Fabric a clip you already shot

The failure mode is muscle memory. Last week you used Fabric. This week you have a take. You open Fabric. You upload a frame from the take as the “still.” You spend 40 credits. You get a 720p talking photo that is not the take. The camera move is gone. The product turn is gone. The cut is gone.

The other failure mode is running Lipsync because it is 4 credits and hoping it will invent a presenter from a product still. It will not start, or it will start on the wrong file. Video-to-lipsync wants video.

A third failure: stacking this pass on a silent generate and calling the result native audio. Native audio is a generate that wrote the track. This is a mouth pass on a track you brought. Fine. Honest. Different bill.

Run the 4-credit row from the catalog when the two files are ready. Confirm video plus audio. If the confirm screen asks for a still and 40 credits, you are on Fabric. Back out.

After the mouth: captions, ratio, label. The 4 credits were the cheap part.

FAQ

Is VEED Lipsync the same as Fabric?

No. Lipsync is video-to-lipsync at 4 credits on a clip plus audio. Fabric 1.0 is image-to-lipsync at 40 credits, 720p, on a still plus audio. Same vendor. Different input. Different sticker.

Do I need a photo for VEED Lipsync?

No. requires_image is false. You need the video and the driving audio. A still of the speaker is a different VEED row.

When is 4 credits the wrong number?

When you do not have the talking take yet. A still plus a track is Fabric or Avatar Pro. A paragraph with no picture is TTS. A new scene with dialogue in one pass is a native-audio generate. Four credits will not do those jobs.