Design each scene like a still, cast an AI presenter, give it a voice, add motion without keyframes, translate the whole thing, and render it. All on one canvas, in the client's brand.
Seven stops, from a blank canvas to a rendered video with a presenter
Here's the walk, in the order you'll do it.
1. Open Video Canvas
it's in the sidebar, under Create
2. Pick a size, build the scenes
every artboard is a scene
3. Cast a presenter
hundreds of AI actors, or one from a photo
4. Give it a voice
a script per scene, lip-synced
5. Add the motion
one click per element, no keyframes
6. Translate the whole video
text and voice, same presenter
7. Render it
into the Video Editor for music and captions
The client wants a Reel with a presenter, and no crew.
Here's where to build it, scene by scene, in their brand.
Step 1 · Open Video Canvas
It's in the sidebar, under Create
Open Create in the left sidebar and click Video Canvas. You land on a start page: your saved projects, a quick avatar video, a presentation from a deck, a custom size, and the community library. The client's brand kit is a panel on the canvas itself, so you pick the client once you're in.
1Open Create
in the left sidebar
2Click Video Canvas
the fifth row — or the Video Canvas card on the dashboard
3Pick a starting point
a saved project, a preset size, or a deck
Video Canvas · the start pageBellhaven Design + Build
Open A Project
browse and resume any saved video project, yours or the team's
Avatar Videos
add your script and quickly generate talking presenters
Avatar Presentation
drop a PDF; every page becomes a scene
Custom Size Video
any size from 360 to 4096 pixels; starts you with three scenes
StyleForge Video Library
community-published starting points
Getting Started
short training videos, from the first scene to the render
Aspect Ratio Presets · a multi-scene project in one click
you design each one like a still, with the same tools as the design canvas, and the scenes play in order.
A presenter without a shoot
an AI actor on the scene, saying your script in a voice you pick, or your own.
No keyframes, ever
click an element, set when it comes in, pick one of twelve motions. That's the whole timeline.
You pick the size, and design each scene like a still.
The client's photos, fonts and colours. The scenes line up along the bottom.
Step 2 · Pick a size, build the scenes
Pick the platform size, then design each scene like a still
Click Instagram Reels and the project opens with scenes already lined up along the bottom. Each one is an artboard: drop in the client's photos, set the headline in their fonts, colour the background from their kit. Add a scene, duplicate one, drag them into order. The durations set themselves from what's on each scene.
4s · the hook: the before-and-after kitchen, headline slides up
SCENE 2
9s · the presenter, green-screen, over the showroom photo
SCENE 3
8s · three benefits, one card each, fade in on a beat
SCENE 4
5s · the offer and the date, zoom in
SCENE 5
4s · the logo, the phone number, the address
The scenes along the bottom · Add Scene · Duplicate · drag to reorder
Scene 14s
Scene 29s
Scene 38s
Scene 45s
Scene 54s
+ Add Scene
On every scene
The client's kit, a panel away
their fonts, colours, logos and saved photos, on every scene
The full design toolset
hundreds of fonts, shapes, effects, colour styles and layers — it's the design canvas underneath
Photos, video, generated pictures
drop in the client's photos, a video background, or generate the picture in the frame
Duplicate, reorder, delete
the scene strip is the running order; drag a card to move it
Durations that set themselves
from the voice and the motion on each scene; nudge them if you want
You cast a presenter, and drop them onto the scene.
Hundreds to choose from, or one made from a photo of the client's own people.
Step 3 · Cast a presenter
Pick an AI presenter, or make one from a photo, and drop them onto the scene
Open the actors panel on the scene where someone speaks. Choose from hundreds of AI presenters, recorded, selfie-style or on green screen, or turn one photo into a presenter and save them to the client's kit so they're the same face next time. Drop a green-screen presenter over your showroom photo and click Composite Greenscreen. One presenter per scene, as many scenes as you like.
Actors · Scene 2RecordedSelfieGreen screen
Maya
Devinon the scene
Priya
Owen
Make one from a photo
the client's own designer as a presenter, saved to their kit
Scene 2 · the presenter over the showroom
2
Composite Greenscreen
the presenter is layered over the background and the elements into one clean scene
Casting
Hundreds of AI presenters
every age, style and setting; recorded, selfie-style, or on green screen, some with more than one look
A presenter from one photo
the client's founder, the designer, the sales lead — turned into a presenter and saved to their kit
Green screen over anything
drop the presenter on your own background and composite it into one layered scene
Variations for ads
one prompt, up to a hundred on-brand takes mixing presenters, products and scenes — save the ones you like
B-roll on any scene
generate the supporting footage right on the scene, from a sentence or a photo
You write what each scene says, and the presenter says it.
Any voice, or your own. Lip-synced. The scene runs as long as the line.
Step 4 · Give it a voice
Write what each scene says, pick a voice, and the presenter lip-syncs it
Open Scene Voice. Type the script for each scene, or let the script writer draft it from what you tell it. Pick a voice, or record thirty seconds of your own and the presenter speaks in it. Click Generate Audio, then Lip-Sync Actors. The presenter's mouth matches the words, and the scene's length follows the voice.
Scene VoiceDraft the scriptRecord with your voice
Scene 1· voice: Maya · warm, American
Your kitchen, done in six weeks, by one team that designs it and builds it.
Scene 2· voice: Maya · warm, Americanlip-synced
Hi, I'm Maya from Bellhaven. Book a spring consultation and the design work is on us.
Scene 3· voice: Maya · warm, American
One quote. One schedule. One team from the first sketch to the last cabinet door.
Generate AudioLip-Sync Actorsthe scene lengthens to fit the line
The voice
A script per scene
type it, or tell the script writer the goal and the tone and it drafts every scene
Any voice, or yours
pick a voice, or record thirty seconds and the presenter speaks in your voice, saved to the kit
Lip-sync that matches
the presenter's mouth follows the words; regenerate any scene on its own
Music and sound
a music track from the library, an upload, or one generated from a sentence, under the whole video
Timing follows the voice
the scene runs as long as the line; the motion lands inside it
You click an element, set its entry, and pick a motion.
Twelve to choose from. No keyframes, no timeline wrangling.
Step 5 · Add the motion
Click an element, set when it comes in, pick one of twelve motions. No keyframes
Click the headline on a scene and press Set Entry: that's the moment it appears. Pick how it arrives, a fade, a slide, a zoom, a bounce. Do the same for the photo and the offer. Stills can drift with a slow Ken Burns move, and any scene can take a video background. That's the whole timeline.
The photo: Fade In at 0s, with a slow Ken Burns drift. The headline: Slide Up at 0.8s. The offer line: Slide Up at 1.4s. Three clicks each.
The scene strip · durations set from the voice and the motion
Scene 14s
Scene 29s
Scene 38s
Scene 45s
Scene 54s
+ Add Scene
Motion, plainly
Set Entry
one click marks when an element appears; that's the timing done
Twelve motions
fade, slide, zoom, bounce and elastic, in and out; pick one per element
Ken Burns on stills
a slow drift across a photo so a still never sits dead
Video backgrounds
drop a clip behind a scene; the elements sit on top
See it on the scene
the motion plays on the artboard as you set it; no separate preview
You duplicate the project, and translate the whole thing.
Text and voice together, same presenter, twenty-nine languages.
Step 6 · Translate the whole video
Pick a language, and every scene's text and voice comes back translated
The client wants the same Reel for their Spanish-speaking customers. Duplicate the project first, because translating replaces the original. Then click Translate All Text & Audio, pick Spanish, and every headline, every line of script and every voice track comes back in Spanish, in the same voice. Run Lip-Sync Actors again and the presenter speaks it. Twenty-nine languages.
Translate All Text & AudioDuplicate project firstSpanish
Scene 2 · Englishoriginal
2
Hi, I'm Maya from Bellhaven. Book a spring consultation and the design work is on us.
Scene 2 · Spanishsame voice · lip-synced
2
Hola, soy Maya de Bellhaven. Reserva una consulta de primavera y el diseño corre por nuestra cuenta.
SpanishFrenchGermanPortugueseItalianJapaneseKoreanHindiArabic+ 20 more
Translating
Text and voice together
every headline and every script line, and the audio regenerated, in one pass
The same voice
the presenter keeps their voice in the new language; run lip-sync again and the mouth matches
Duplicate first
translating replaces the original project, so duplicate it and keep the English one
The whole project at once
one click covers every scene and every presenter, not one clip at a time
Twenty-nine languages
one build, as many markets as the client sells in
You click Render, and finish it in the Video Editor.
Music, captions, A/B versions, and the MP4.
Step 7 · Render it
Click Render, and the scenes open in the Video Editor for music, captions and the MP4
When the scenes are right, click the purple Render button. Your scenes open in the Video Editor, where you add the music track, turn on captions, make A/B versions by swapping the hook or a photo, and export a high-resolution MP4. Render a single scene if you only changed one.
Spring kitchen offer · ReelRender Single SceneRender
Scene 14s
Scene 29s
Scene 38s
Scene 45s
Scene 54s
+ Add Scene
In the Video Editor · what happens next
Pick the scene variations
the takes you want in the cut
Music and narration
a track under the whole video
Captions
generated from the voice, styled to the brand
A/B versions
swap the hook or a photo; two cuts from one build
A high-res MP4
fast render, ready to post
Rendering
One button
the purple Render button hands every scene to the Video Editor
Render one scene
changed a line on Scene 2? Render just that one
It lands in the client's library
the finished video sits with the client's other videos, shared with the team
Come back and change it
the project stays on the canvas; edit a scene, render again
Two videos from one build
the English Reel and the Spanish one, rendered from the same scenes
Or you drop in a deck, and every page becomes a scene.
The narration drafted for you. Add a presenter, and it talks.
Or · Start from a deck
Or drop in the client's deck, and every page becomes a scene with a presenter
Pick Avatar Presentation on the start page and drop in a PDF, the client's pitch deck or slides exported from their presentation tool. Each page becomes a scene with the slide as the background. The script writer reads every page and drafts the narration. Add a presenter, lip-sync, and a twenty-page deck is a talking presentation in minutes.
Avatar PresentationMax 30 pages
Drop a PDF document here
bellhaven-spring-pitch.pdf · 14 pages
The script writer
“A friendly walkthrough of the spring offer for homeowners, two sentences a page, warm and plain.”
A scene per page
the slide is the background; the layout is yours to touch up
Narration drafted for you
the script writer reads every page; edit any line
One presenter, every page
lip-synced to the script, in the voice you pick
Then translate it
the same deck in Spanish for the same money
Good to know
The presenters and voices are AI. Make one from a photo of a real person only with their say-so.
One presenter per scene, as many scenes as you like.
Translating replaces the original project. Duplicate it first and keep both.
The render finishes in the Video Editor. That's where music, captions, A/B versions and the MP4 live, and where the finished video is filed with the client's other videos.
Saved video templates are on the start page as coming soon. The community library and your own projects are live.
The client on this page is fictional. The real canvas fills with the client's own kit.
The walk loops — next month's offer opens this project, swaps the photos and the script, and renders again.