Skip to content
Aardvark Studio
Get the Core

VisualBook

Contact for pricing

Produces

A long-form video storybook

VisualBook turns a book the studio has already read into a long-form video storybook — not a trailer, not an audiobook with a cover on it, but the work itself, seen and heard, scene by scene. It casts each character to a visual identity that holds from frame to frame, breaks scenes into beats and beats into shots, and assembles the approved frames against the narration until a chapter is something you can sit and watch.

Requires Aardvark Studio Core (free) · Release 1.0.0

Who it is for
Anyone turning a title they own into screen work — a trailer, a chapter, or the whole book — without a studio behind them, and anyone who wants to work in a long-form visual format before it has a shelf to sit on.
What it takes from the Core
An ingested book: the narration aligned to the prose, the scene and beat breakdown, and the cast, props and locations already extracted and approved.
What you have to run
The Aardvark Studio server · A ComfyUI render host with a GPU · An Anthropic API key

The change

Before, and after

Today, without it

  • A folder of AI images that do not agree on what the character looks like.
  • A shot list in a spreadsheet, and no link from a shot back to the sentence it came from.
  • Re-prompting from scratch every time a scene changes.
  • An edit timeline you have to sync to the narration by hand.

With VisualBook

  • One casting portfolio per character, approved once, reused everywhere.
  • Shots derived from the beats of the actual prose, with coverage you can read off a matrix.
  • Rework is a decision on a run, not a new prompt in a browser tab.
  • Assembly times the approved frames to the audiobook and writes the chapter markers itself.

How it goes

A book has two finished forms today: one you read and one you listen to. This is the argument for a third — the whole work, rendered and timed to its own narration, watchable end to end. Nobody has a name for that shelf yet, which is the interesting part.

Nothing here starts from a blank prompt. The book has already been read: the narration is aligned to the prose, the chapters are broken into scenes and beats, and the cast, props and locations have been pulled out and approved. A shot exists because a beat exists, and a beat exists because a sentence does.

So the work is casting and judgement, in that order. Cast each character once and keep the images that are right. Let the storyboard propose shots per beat and approve the ones that read. Preview a stretch as often as you like while you are still shaping it. Then assemble — every approved frame, against the full audiobook, with the chapter markers already in place.

Where it stops for you

3 gates, and who holds each

  1. 1

    Casting steps

    Casting Director

    Each recipe step lands for review; the curation gate refuses to close until you have kept enough images.

  2. 2

    Storyboard

    Reviewer

    Frames are approved, reworked or rejected per beat. A rejected frame provably never reaches the cut.

  3. 3

    Final assembly

    Reviewer

    The finished render goes through the same sign-off as every stage before it.

Casting that holds a face

A character is cast once, to an identity root, and every later image is built from it. Selection is structural — an image is in the portfolio because a person kept it, not because a flag was set — so a rejected variant provably never reaches training or the screen.

  • An identity root per character, then variants and in-context frames
  • A curation gate that refuses to close under-curated
  • LoRA training where identity has to hold across a whole book
Casting Stage — an identity root and its approved variants

Props, sets and the art department

Props and locations are resolved from the same breakdown the cast came from, so a scene knows what is in it. Sets are placed; props are derived from where they are mentioned rather than pinned to a scene by hand.

  • Prop and set resolution over the ingested book
  • Facet extraction — the burned house, the changed wardrobe, the new scar
Storyboard Frames — one beat, its shots, and their review state

A storyboard made of beats

Each beat becomes shots, and each shot is reviewed. Shot readiness and the coverage matrix answer the question a spreadsheet cannot — what is not ready, and which scenes have nothing to show for them.

  • Frames per beat, reviewed individually
  • Coverage matrix across the whole book
  • Manual frames for intro, chapter cards and credits
Coverage Matrix — what is ready and what is not

Assembly, timed to the narration

Preview a stretch of frames while you are still shaping a scene, as often as you like — segment previews are disposable and need no sign-off. When it is ready, the final assembly renders every approved frame against the full audiobook with gentle motion, and turns your chapter markers into jump points.

  • Segment preview by frame, beat or scene id
  • Final cut timed to the audiobook
  • Chapter markers become jump points in the finished file

Sample output

What comes out of it

These are slots. No sample from a real title has been cleared for publication yet, and a mock-up of one would be the exact thing this page is trying not to do.

Casting Stage — an identity root and its approved variants
Storyboard Frames — one beat, its shots, and their review state
Coverage Matrix — what is ready and what is not

What it needs

Be clear about the hardware

  • required

    The Aardvark Studio server

    One container, deployed once. Every product needs it.

  • required

    A ComfyUI render host with a GPU

    This is the honest one. Casting, storyboard frames and assembly all render on a GPU you point the server at. It can be your own workstation on the same network; it is not included and it is not in the cloud unless you put it there.

  • required

    An Anthropic API key

    Breakdown, casting prompts and captioning all call it.

  • optional

    A LoRA training host

    Needed where a character's identity has to hold across many frames. Without it you can still cast, storyboard and assemble.

Roles it adds

Who can do what

  • Casting Director

    Owns the cast and the curation gate.

  • Prop Master

    Owns props and their continuity across scenes.

  • Set Designer

    Owns locations and their placement.

Worth knowing

VisualBook's three roles open Core workspaces rather than a workspace of their own, so granting the licence from a checkout assigns no roles by itself — the licence still decides what your organisation may assign.

Worth knowing

Its licence is enforced by visibility: without it the Pre-Production, Production and Post-Production surfaces are absent. Its HTTP routes are not seat-gated today.

Worth knowing

A full render is not instant and we publish no timing figure, because the honest one depends entirely on the GPU you point it at.

With the others

VisualBook on its own is complete

Nothing below is required. Each line is what happens if you already own the other one.

+ Loomwright

Film the book you wrote here. The published manuscript is the same source an ingested ePub would be.

A finished book

+ LookBook

Cast the whole book in a published style — one choice, made at Pre-Production ▸ Cinematography.

A published casting style

Questions

About VisualBook specifically

Ready to try it?

The Core is free and does not need this product. Start there, then ask us about a licence.

Do I need my own GPU?
Yes — or access to one. Rendering happens on a ComfyUI host the server calls over the network. That can be a workstation in the same room. There is no bundled cloud renderer.
Can I use it on a book I did not write?
Technically yes; legally that is yours to clear. The studio does not check rights in the material you feed it, and we are not going to pretend it does.
How long does a book take to render?
We do not publish a figure. It depends on your GPU, your recipe and how many frames survive review — and a number measured on our hardware would tell you nothing about yours.
Can I put my own title cards in?
Yes. Design an intro, chapter-card or credits frame yourself and the assembly uses it in place of the automatic slide.