Back to Playlist · VidFarm Walkthrough Tutorial
Part 22 of 33Edit Demo — Animate Static Book8 min read
Edit demo #5 — animate a static book

Turn a flat book cover into a living, moving scene

Two ways to bring a still image to life: a free Ken-Burns push you can run on any tier, and an AI image-to-video move that invents the motion around your frame. Keep the same character in every shot.

Copies the whole guide — prompts, next steps & all — for your AI agent
Watch this chapter
Start with one flat book-cover PNG, add a slow push-in, then generate real motion from the same frame — and post the finished recap.

Key takeaways

  • The cheapest way to animate a still is a Ken-Burns pan/zoom on an image layer — free, no AI key, works on any tier.
  • For real motion, image-to-video uses your still as the first frame and the model invents the rest.
  • You animate by chatting — "animate all the images, vary the motion" — and the copilot applies it live.
  • Keep a recurring face on-model with a character sprite card reference (see the companion guide).

The idea: a still is just the first frame

A book cover, an illustrated page, a product shot — any flat image can become a scene. There are two levers, and you pick by budget. The free one is Ken Burns: a slow pan or zoom on an image layer that fakes camera movement with zero generation. The richer one is image-to-video: you hand the model your still as the first frame and describe the motion, and it animates outward from there.

Everything below happens in the Trackpad Editor by chatting with the AI copilot. Forking is automatic — the first edit mints your private copy, and there's no save button. Let's build a short book recap.

A hand-drawn six-panel storyboard sketch on cream paper, three columns by two rows. Panel 1: a closed storybook on a table with sparkles rising. Panel 2: the pages fly open and a little fox in a scarf steps out. Panel 3: the fox walks a winding path over rolling hills under a smiling sun. Panel 4: the fox looks up at a tall mysterious forest gate. Panel 5: the fox crosses a small wooden bridge over a river. Panel 6: the fox stands on a hilltop at sunset, content. Each panel a distinct scene telling one legible little story.
A book is just a stack of stills waiting to move. Sketch the beats first — one distinct scene per panel — then bring each still to life with a Ken-Burns push or a generated motion shot.

Manual walkthroughDo it yourself (the web UI)

  1. Fork a book-recap template and get your stills in

    Pick a book-recap or listicle format on Discover and open it in the editor. Then bring your images into My Files — the cover and any page illustrations. In chat: "Use my book cover and these page images from My Files as the scenes." The copilot runs browse_files, reads each file (its view_url is the ready media URL), and swaps them onto the timeline with set_layer_media or add_layer kind=image.

    Open the editor

  2. The free move — animate every still with Ken Burns

    Chat: "Animate all the images — slow Ken Burns, vary the motion between scenes." The copilot applies a ken_burns preset to every image layer (zoom-in, then pan-left, then zoom-out…) so no two adjacent stills move the same way. This costs nothing to generate and runs on any tier — it's just CSS motion on an image. You can also set it by hand in the inspector's Animate (Ken Burns) dropdown.

  3. The richer move — image-to-video from the same frame

    Want the cover to actually breathe? Chat: "Animate this book cover — slow push-in, the title glints, embers drift across the frame. 5 seconds." The copilot fires generate_layer with media_type: video, passing your still as the conditioning frame (the still's URL as input_references, or a frame_images entry with frame_type: "first_frame"). AI video costs real money, so the copilot asks permission first before it runs.

  4. Add narration, a music bed, and captions

    Chat: "Add a calm narrator reading these lessons, and a low lofi music bed under it," then "Add karaoke captions recapping the book's core lessons." The copilot lays a narration track via /audio/speech (~1.0 volume) and a real music file underneath (~0.1–0.2). There is no music-generation primitive — the bed comes from an owned or uploaded file, never faked by duplicating the voice.

  5. Render, then approve into a shareable post

    Review the live preview. When it's right, say "render it" — the copilot never auto-renders. The finished MP4 lands as a shareable post on its approved page with a copyable caption, a Download button, and Schedule. It also shows up in your approved library .

The Trackpad Editor inspector with the Animate (Ken Burns) dropdown open on a selected book-cover image layer, showing slow-zoom and pan presets, with the cover previewed on the canvas.
Step 2 — the free path. The Animate (Ken Burns) dropdown adds a slow pan or zoom to any image layer, no AI key required.
The editor generate panel for image-to-video: a first-frame slot holding the book-cover still, a motion prompt reading 'slow push-in, the title glints, embers drift', a 5-second duration field, and a cost-confirmation prompt.
Step 3 — the AI path. Your still becomes the first frame; the model invents the motion around it. The copilot confirms the cost before generating.
What it costs The Ken-Burns push is free and runs even on the free tier — it's pure CSS motion, not generation. Image-to-video is a real AI-video job billed to your keys, so start with Ken Burns and only spend on generation for the hero shot that needs it.
Heads up On a free NVIDIA "lite" key, the AI image-to-video path is blocked (it throws LITE_TEXT_ONLY_MEDIA_ERROR) — but Ken-Burns animation of a static image still works, because it's not generation. To generate motion you need a BYOK OpenAI, Gemini, or OpenRouter key.

Keep the same character in every shot

If your book has a recurring character — a mascot narrator, an illustrated protagonist — you don't want their face drifting between scenes. Lock them once into a character sprite card and pass that same reference into every generation. That's the whole trick to on-model consistency, and it's covered in its own guide.

Character consistency across shots

Note Ken Burns is images-only — it never applies to video, text, or audio layers. Vary the preset across adjacent stills so the motion doesn't visibly repeat, and keep the intensity gentle (0.04–0.5) so the frame never crops past its edges.

Where to go next

A static image is the simplest thing to animate. Once you're comfortable, string several animated scenes into a full narrative with recurring characters — that's the AI drama series.

Animate your first still now

Open the editor, drop in a cover, and ask for a slow Ken-Burns push. It costs nothing — and when you want real motion, the same frame becomes your first frame.

Open the editor