How to Create Video with AI Space Note: Four Methods on One Canvas
AI9_Studio ·
How to Create Video with AI Space Note: Four Methods on One Canvas Most AI video tools give you a single box: type a prompt, wait, download. AI Space Note works differently. It is an infinite canvas…
How to Create Video with AI Space Note: Four Methods on One Canvas
Most AI video tools give you a single box: type a prompt, wait, download. AI Space Note works differently. It is an infinite canvas of note cards — text, images, video, audio, files — and every card on that canvas can become the source of your next video. Your reference photo, your script, your voiceover and the finished clip all live side by side, connected by lines that show what came from what.
This tutorial walks through the four ways to create video in AI Space Note, from a single click on an image to a full multi-scene workflow planned by the built-in AI agent.
Before you start
Sign in at ai9app.io
and open AI Space Note from the menu.
Your first three notebooks are free; after that each new notebook costs 50 credits, and every notebook holds unlimited cards.
Add something to the canvas first. Drag in an image or video, paste a YouTube link, upload from My Stuff, or double-click empty space to write a text card. Video needs a starting point, and in AI Space Note that starting point is always a card.
Method 1: Animate — turn any image into a moving shot
This is the fastest route from a still to a clip.
Hover any image card and click Animate. A Generate window opens with the image already attached as the first frame. Write what should move — camera, subject, lighting, mood — then set the duration on the slider, pick an aspect ratio and a resolution, and press Generate.
The default model is Seedance 2.0, which handles 4–15 seconds and understands camera language well (“slow orbit”, “push in”, “handheld”). Open the model picker to switch: Seedance 2.5 goes up to 30 seconds, Google Veo 3.1 renders 8-second shots up to 4K, and the Kling and MiniMax models each bring their own motion character.
The result lands as a new video card connected to the original image, so you can see the lineage of every shot later.
Method 2: Generate — a prompt, a model, a duration
You do not need an image to start. Use the toolbar to add a video card, open Generate video, and write the shot from scratch.
The model picker groups by provider so you can compare price and capability in one place. Two settings deserve your attention before you press Generate.
Duration
Models with a range show a one-second slider: Seedance 2.0 gives 4–15s, Seedance 2.5 up to 30s, Omni 3–10s. Fixed-length models show a dropdown instead — Veo 3.1, for example, is always 8 seconds.
Aspect ratio
Pick 9:16 before you generate for Reels, Shorts and TikTok. Re-cropping afterwards costs you framing you already paid for.
The four-part prompt
A prompt that produces a usable shot names four things: subject, style, camera movement, and how the shot ends.
“A white ceramic bottle on a stone plinth, clean studio lighting, slow 180-degree orbit, ending on a front-facing close-up” beats “a nice product video” every time.
Method 3: Merge — combine several cards into one shot
Merge is what a canvas is actually for. Click Merge on any card, then click other cards on the board to attach them as references, and choose Merge to video.
What you can attach depends on the model you pick, and the interface only lets you select what will actually work:
Seedance 2.0 accepts up to 9 reference images, plus a reference video and a reference audio track
Seedance 2.5 accepts up to 30 reference images
Kling and MiniMax models accept 1–4 images
Text cards can be attached too — up to three — and their text is appended to the prompt
This is how you keep a character, a product or an art direction consistent across shots: attach the same character sheet to every scene. If you use one recipe repeatedly, save it as a merge preset and reuse it with one click.
Method 4: The AI agent — describe the result, approve the plan
The first three methods make one video at a time. The agent makes a whole set.
Open the AI Assistant from the toolbar and switch to Computer mode. Describe the outcome in plain language — “make a 6-scene product story for this bottle, 9:16, 8 seconds per scene, with the same model in every shot” — and press send. Pick the cards you want it to work from first, or upload a file, and the agent builds the answer from those instead of guessing from the board.
The agent replies with a plan, not a video. Each step shows the chosen model, the prompt it wrote, the duration, the references and the estimated credit cost, with a total at the bottom. Every field is editable: change a model, rewrite a prompt, drag a duration, add a reference image, delete a step. Nothing is generated and nothing is charged until you press Approve & generate.
Three things make the agent worth using for multi-scene work.
Character consistency
The agent can generate one base image, then reuse it automatically as the reference for every later scene, so the same face and outfit survive across shots.
Mixed output
A single plan can write text cards, generate images, animate them into video, and record a multi-speaker podcast track for narration — all in the same approved run.
Supervised execution
Steps run in parallel with automatic retries, and a Stop button is available throughout. Interrupted runs are labelled rather than silently resumed.
Finish the video: captions and the Video Editor
Two steps close the loop. On any video card, open Subtitle / Notes to generate captions in your chosen caption language and drop them onto the board as a text card you can edit.
When the shots are ready, use Import to Video Editor. Pick your video and audio cards, set the order and assign each to the main track, an overlay track or the audio track, and the whole sequence opens in the AI9app Video Editor for trimming, captions and final export.
Frequently asked questions
Can I generate video without an image?
Yes. Method 2 is prompt-only; Methods 1 and 3 use cards as references.
How long can a video be?
It depends on the model — 30 seconds is the current single-shot maximum, with Seedance 2.5. For anything longer, generate scenes and assemble them in the Video Editor.
Will the agent spend credits on its own?
No. Plans are proposals; generation starts only when you approve, and the estimated cost is shown before you do.
What happens if I leave the page mid-generation?
Runs continue on the server and reattach to their cards when you come back.
Start your first video
Open AI Space Note, drop one image on the canvas, and press Animate. That is the whole first lesson — everything else on this page is a way to do the same thing at greater scale.
Start free at ai9app.io/ai-space-note