BGs

VIDEO STUDIO

Accounts are handed out by invitation. There is nothing to sign up for here.

…

↑ ↓ Enter · press / to search+ New channel

Channels

Everything that is the same in every video of a channel.

Videos

What is being made, what waits for you, and everything you have made.

All videos
WhenVideoStatus
← Videos

Add a presenter

Already have a finished video — footage, edit and voice-over done elsewhere? Drop it here and one of your cast will present it. They lip-sync to the voice-over that is already in the file: nothing is re-voiced and the sound is left exactly as you mixed it.

The finished video

Stays on this machine. The picture is rebuilt once, the sound is copied across untouched.

Who presents it
How they appear

They come in where the voice-over pauses, opening and closing the video — the same rhythm as your own videos: about one appearance a minute and a seventh of the running time.

← Videos

Scene editor

Change a scene here, then make just that scene again. Nothing is generated until you press a button.

Video
ScenePreviewKind · length What it showsEnginesState

Your cast

The people who can front a video. Each one needs a face HeyGen can drive and a voice. Nothing is generated until you press a button.

Bodycam

Each episode gets its own cast: faces, wardrobe passports and voices. Review them here before the expensive stages run — listen to a voice, change it, and the video will use your pick.

No bodycam videos yet.

Setup

Studio-wide settings — the same for every channel. Each card saves on its own; the choices at the top save the moment you pick them.

How videos are made
Who writes the script
Claude writes the script and plans the scenes.
What makes the clips
Veo 3.1 Fast on snapgen.ai — the clip engine
Every AI channel shoots its moving clips on Veo 3.1 Fast. When Veo refuses a clip or runs out of credits, Grok on kie.ai makes it instead and the video's card says so. Channels on real footage buy no clips.
Sharper wide shots
Nature and wide shots get their detail spread over the whole frame, so the normal setting turns brush and grass into mush. Close-ups, hands and screens look the same either way.
The person on camera …

Signing in spends the credits in your HeyGen plan. An API key would bill per video instead — that is why this is a sign-in and not a key.

YouTube footage

One sign-in for the whole studio. Channels on real footage download archive clips with it.

Keys

They stay on the server — this page can put one in, but there is no way to read one back out, not even for you. A box left blank keeps the key that is already there.

Editing a host
New host
Portrait
no portrait yet
Age
Gender
Who they are
How close
Behind them
Where they are — leave empty and the room below decides
Reference picture — optional: the place, or a shot whose angle and pose to copy
none
The room

Close and blurred is what the talking presenter needs — the bigger the face, the cleaner his speech comes out. The wider frames keep the hands out of shot and empty: this stays a portrait, not somebody showing you a thing. The room runs from their own cluttered workroom to a well-off house: clutter makes a person believable, but on a money, pensions or law channel it reads as poverty rather than truth.

One picture, about 20 seconds. This one costs money.

A new portrait is re-sent to HeyGen as a new face: press Publish (or Save). If you do not, the next video sends it before filming — the old face is never used again.

Name
Voice
Voice engine · how it reads
Character · optional, blank = voice defaults
Clone a new voice · from your recordings

Clean quiet recordings only — the clone copies the recording quality, not just the voice. Uses a voice slot.

On camera · how much of them moves

Avatar III is the cheap one and suits most faces. If it looks wrong on this one, move them up — only they get pricier.

Notes · optional — how they sound
Fronts these channels