---
title: A storyboard from a script with a consistent cast
display: 'A *storyboard* from a script with a consistent cast'
nav_title: Storyboard from a script
summary:
  - "A six-panel pencil storyboard with the same cast, lantern and screen direction in every shot, captions set in code."
  - "Build two reference sheets and one shared prompt block first; each panel only adds its shot line."
  - "Nine image runs and about 3 minutes of waiting; once the sheets exist, each new panel is one run."
description: Turn a short script into six numbered storyboard panels with the same cast, props and screen direction, with the captions set in code.
section: playbooks
order: 110
type: playbook
topic: Film and animation
level: Build
roles: [artist, art-director]
time: { read: 10, do: 30, wait: 5 }
needs: ["Any plan with credits", "A script broken into shots", "Python with Pillow, ImageMagick or any layout tool for the captions"]
last_verified: 2026-10-16
models_checked: ["GPT Image 2.5 Sunburst"]
hero: playbooks/storyboard-from-script/quillfern-lantern-storyboard-board.webp
related: [character-reference-sheet, long-video-from-guided-frames]
pdf: true
run_in: [model_openai-gpt-image-2-5-sunburst]
# the use case made, on scenario.com/explorations, and the platform workflows that run it
made_with:
  - { slug: previs-to-picture, title: "Previs to Picture" }
  - { slug: the-last-lantern, title: "The Last Lantern" }
run_with:
  - { id: wflow_cinematic-grid, name: "Cinematic Grid" }
  - { id: wflow_next-scene, name: "Next Scene" }
thumb: playbooks/storyboard-from-script/thumb.webp
video: tutorials/storyboard-from-script.mp4
chapters:
  - { at: 0, title: "The storyboard to make", image: "/media/images/playbooks/storyboard-from-script/ch-01-the-storyboard-to-make.webp" }
  - { at: 19, title: "One line per shot", image: "/media/images/playbooks/storyboard-from-script/ch-02-one-line-per-shot.webp" }
  - { at: 38, title: "A cast sheet and a prop sheet", image: "/media/images/playbooks/storyboard-from-script/ch-03-a-cast-sheet-and-a-prop-sheet.webp" }
  - { at: 60, title: "A shared block, unchanged", image: "/media/images/playbooks/storyboard-from-script/ch-04-a-shared-block-unchanged.webp" }
  - { at: 80, title: "Add one shot line", image: "/media/images/playbooks/storyboard-from-script/ch-05-add-one-shot-line.webp" }
  - { at: 102, title: "Describe only who’s there", image: "/media/images/playbooks/storyboard-from-script/ch-06-describe-only-who-s-there.webp" }
  - { at: 122, title: "Captions set in code", image: "/media/images/playbooks/storyboard-from-script/ch-07-captions-set-in-code.webp" }
  - { at: 135, title: "The board, animated", image: "/media/images/playbooks/storyboard-from-script/ch-08-the-board-animated.webp" }
  - { at: 154.5, title: "Recap", image: "/media/images/playbooks/storyboard-from-script/ch-09-recap.webp" }
steps:
  - title: "One line per shot"
    text: "Create › Text opens Scenario LLM: paste the instruction, and add the script under Text Inputs. The model drafts one line per shot with who is on the left and right; you fix the cuts by hand."
    prompt: |-
      Break this script into a storyboard shot list. One line per shot: number, shot size, camera angle, the action in one sentence, duration in seconds (2 to 5). Keep screen direction consistent and say who is on the left and who is on the right in every shot. Script: [paste the script]
  - title: "A cast sheet and a prop sheet"
    text: "Before any panel, make two color sheets: Isa and Master Oren in three views with expressions, and a prop and location sheet. The lantern is described part by part (round amber globe, exactly three brass bands, ring handle), because counting details keeps it the same later."
    prompt: |-
      Character reference sheet for an animated feature film. The reference image is Isa: keep her face, freckles, wavy auburn hair tied back, brass goggles on her forehead, cream shirt with rolled sleeves, soot-stained brown leather apron, olive trousers and worn brown boots exactly as shown. Plain warm light-grey background, no text, no labels.

      Two characters, each shown full body in three views side by side: front, three-quarter, back. Same scale, standing on one faint ground line, empty hands.

      Left group, ISA: the 12-year-old glassblower's apprentice from the reference image.

      Right group, MASTER OREN: a stout elderly glassblower, a head and a half taller than Isa, bald with a thick white walrus mustache, round leather-rimmed spectacles, rolled shirt sleeves showing strong forearms, a long burn-marked dark leather apron over a faded blue shirt, grey trousers, heavy boots.

      Bottom row, small: four head close-ups of Isa (proud smile, surprised, worried, determined) and two of Oren (warm smile, laughing).

      Style: prestige painterly 3D feature animation, hand-painted textures, expressive appealing shapes, soft volumetric light, rich color. Even soft studio light. Clean gutters between figures.
  - title: "A shared block, unchanged"
    text: "Create › Images › GPT Image 2.5 Sunburst. Reference Images › Select from Library › the cast sheet, then the prop sheet; Dimensions 1536x864, Quality high; then paste the shared block that starts every panel prompt."
    prompt: |-
      Storyboard panel for an animated feature film. Loose graphite pencil sketch on plain white paper, quick confident lines, light grey shading only, no color, no text, no numbers, no arrows, no panel border. Image 1 is the cast sheet: Isa, a 12-year-old girl with wavy hair tied back, goggles on her forehead, rolled shirt sleeves, a long leather apron, trousers and boots; Master Oren, stout and a head and a half taller, bald, thick white walrus mustache, round spectacles, rolled sleeves, long leather apron. Match their proportions, faces, hair and costume shapes. Image 2 is the prop and location sheet: Isa's round glass globe lantern with three brass bands and a ring handle on top, the single stone arch of the Great Bridge hung with lanterns, Oren's canal-side workshop with its round furnace and stone steps, and the canal between tall leaning houses. In every exterior shot the canal flows from left to right and the Great Bridge is on the right side of the frame.
  - title: "Add one shot line"
    text: "Under the block, add one shot line. Panel 5: over Isa's shoulder, the lantern drifts left to right toward the bridge. All six panels ran at once."
    prompt: |-
      Shot: 5. OVER THE SHOULDER, behind Isa standing in the punt. The lantern floats away from her on the current, left to right, toward the Great Bridge on the right side of the frame, getting smaller. Isa reaches one arm out after it.
  - title: "Describe only who’s there"
    text: "Panel 6 came back with Oren rowing beside her, because the block described both characters. The retake keeps only Isa in the block and says she stands under the only Great Bridge in the story."
    prompt: |-
      Storyboard panel for an animated feature film. Loose graphite pencil sketch on plain white paper, quick confident lines, light grey shading only, no color, no text, no numbers, no arrows, no panel border. Image 1 is the cast sheet: Isa, a 12-year-old girl with wavy hair tied back, goggles on her forehead, rolled shirt sleeves, a long leather apron, trousers and boots. Match her proportions, face, hair and costume shapes. Image 2 is the prop and location sheet: Isa's round glass globe lantern with three brass bands and a ring handle on top, and the single stone arch of the Great Bridge hung with lanterns.

      Shot: 6. WIDE, dusk, under the arch of the Great Bridge. Isa stands alone in the punt holding the lantern high in both hands, its warm light filling the underside of the arch and reflecting on the water; the hanging lanterns of the parapet glow above. She is UNDER the only Great Bridge in the story: no other bridge, arch or viaduct appears anywhere in the frame, and the canal beyond the arch opens onto calm water and rooftops only.
  - title: "Captions set in code"
    text: "Panel numbers, shot sizes and seconds never go to the model. A short Python script sets them under each panel from the same shot list, then lays out the board."
  - title: "Idea → Storyboard → Animated Video"
    text: "To run this every time, build it as a workflow, or open the Idea → Storyboard → Animated Video template, which goes on to video."
agent_steps:
  - title: "Connect the Scenario MCP"
    text: "Connect your agent and sign in once at [mcp.scenario.com](https://mcp.scenario.com). Then tell it which Scenario team and project to use, so every file lands there."
  - title: "Give the agent the files"
    text: "The script, already broken into scenes, the approved portrait of Isa and a city mood image for the prop and location sheet. A local agent also needs Python with Pillow (or ImageMagick) to set the captions and lay out the board."
  - title: "Brief the agent: shot list and sheets"
    text: "The agent drafts the shot list itself; you fix the cuts before any panel. Then it makes the two color sheets the panels will follow."
    prompt: |-
      First, break the script into a shot list with this instruction, word for word, and show it to me as a table. Do not draw anything until I approve it:
      Break this script into a storyboard shot list. One line per shot: number, shot size, camera angle, the action in one sentence, duration in seconds (2 to 5). Keep screen direction consistent and say who is on the left and who is on the right in every shot. Script: [paste the script]

      When I approve the list, make the cast sheet with GPT Image 2.5 Sunburst (model_openai-gpt-image-2-5-sunburst), width 2048, height 1152, quality "high", with Isa's approved portrait as the only reference, in referenceImages as an array. Prompt, word for word:
      Character reference sheet for an animated feature film. The reference image is Isa: keep her face, freckles, wavy auburn hair tied back, brass goggles on her forehead, cream shirt with rolled sleeves, soot-stained brown leather apron, olive trousers and worn brown boots exactly as shown. Plain warm light-grey background, no text, no labels.

      Two characters, each shown full body in three views side by side: front, three-quarter, back. Same scale, standing on one faint ground line, empty hands.

      Left group, ISA: the 12-year-old glassblower's apprentice from the reference image.

      Right group, MASTER OREN: a stout elderly glassblower, a head and a half taller than Isa, bald with a thick white walrus mustache, round leather-rimmed spectacles, rolled shirt sleeves showing strong forearms, a long burn-marked dark leather apron over a faded blue shirt, grey trousers, heavy boots.

      Bottom row, small: four head close-ups of Isa (proud smile, surprised, worried, determined) and two of Oren (warm smile, laughing).

      Style: prestige painterly 3D feature animation, hand-painted textures, expressive appealing shapes, soft volumetric light, rich color. Even soft studio light. Clean gutters between figures.

      Then make a prop and location sheet with the same settings and the city mood image as the reference: Isa's round amber glass globe lantern with exactly three brass bands and a ring handle, the single stone arch of the Great Bridge hung with lanterns, Oren's canal-side workshop with its round furnace and stone steps, and the canal between tall leaning houses. Quote the cost with a dry run first and wait for my go.
  - title: "Brief the agent: the panels"
    text: "One shared block starts every panel prompt, unchanged, and one shot line goes under it. Keep the block in a file and have the script append each line, so the agent cannot paraphrase it between panels."
    prompt: |-
      Make one panel per approved shot with GPT Image 2.5 Sunburst (model_openai-gpt-image-2-5-sunburst): referenceImages as an array, the cast sheet first and the prop and location sheet second; width 1536, height 864, quality "high". Save this block to a file and start every panel prompt with it, word for word, then add the shot line on a new paragraph:
      Storyboard panel for an animated feature film. Loose graphite pencil sketch on plain white paper, quick confident lines, light grey shading only, no color, no text, no numbers, no arrows, no panel border. Image 1 is the cast sheet: Isa, a 12-year-old girl with wavy hair tied back, goggles on her forehead, rolled shirt sleeves, a long leather apron, trousers and boots; Master Oren, stout and a head and a half taller, bald, thick white walrus mustache, round spectacles, rolled sleeves, long leather apron. Match their proportions, faces, hair and costume shapes. Image 2 is the prop and location sheet: Isa's round glass globe lantern with three brass bands and a ring handle on top, the single stone arch of the Great Bridge hung with lanterns, Oren's canal-side workshop with its round furnace and stone steps, and the canal between tall leaning houses. In every exterior shot the canal flows from left to right and the Great Bridge is on the right side of the frame.

      For example, panel 5's shot line, word for word:
      Shot: 5. OVER THE SHOULDER, behind Isa standing in the punt. The lantern floats away from her on the current, left to right, toward the Great Bridge on the right side of the frame, getting smaller. Isa reaches one arm out after it.

      Start all panels with wait: false and collect them with one jobs_wait. Quote the total first and wait for my go. Never put panel numbers, shot sizes or seconds in a prompt.
  - title: "Check each panel against its shot line"
    text: "Accept when both characters keep their silhouettes, the lantern keeps its globe, bands and ring, the canal flows left to right with the Great Bridge on the right, and each panel shows exactly who its shot line names. Reject a character the script does not put there, a duplicated landmark or written captions. In the run, panel 6 came back with Oren rowing beside her, because the block described both characters."
    prompt: |-
      For each panel, tell me who is in it, which way the canal flows, where the Great Bridge is, and whether the lantern has its three bands and ring. Redo only the panels that fail. For panel 6, use this retake prompt, word for word, with the same references and settings:
      Storyboard panel for an animated feature film. Loose graphite pencil sketch on plain white paper, quick confident lines, light grey shading only, no color, no text, no numbers, no arrows, no panel border. Image 1 is the cast sheet: Isa, a 12-year-old girl with wavy hair tied back, goggles on her forehead, rolled shirt sleeves, a long leather apron, trousers and boots. Match her proportions, face, hair and costume shapes. Image 2 is the prop and location sheet: Isa's round glass globe lantern with three brass bands and a ring handle on top, and the single stone arch of the Great Bridge hung with lanterns.

      Shot: 6. WIDE, dusk, under the arch of the Great Bridge. Isa stands alone in the punt holding the lantern high in both hands, its warm light filling the underside of the arch and reflecting on the water; the hanging lanterns of the parapet glow above. She is UNDER the only Great Bridge in the story: no other bridge, arch or viaduct appears anywhere in the frame, and the canal beyond the arch opens onto calm water and rooftops only.
  - title: "Caption, lay out, file and report"
    prompt: |-
      Set each panel's number, shot size, action and seconds under it in code, from the same shot list, and lay out the board as one image. Create a collection named after the story, add the two sheets and the kept panels, and report the model, every asset ID, which panels you redid and the total cost spent.
updated: 2026-10-16
canonical: https://learn.scenario.com/playbooks/storyboard-from-script
---

Last updated: 2026-10-16

# A storyboard from a script with a consistent cast

> Turn a short script into six numbered storyboard panels with the same cast, props and screen direction, with the captions set in code.

In short:

- A six-panel pencil storyboard with the same cast, lantern and screen direction in every shot, captions set in code.
- Build two reference sheets and one shared prompt block first; each panel only adds its shot line.
- Nine image runs and about 3 minutes of waiting; once the sheets exist, each new panel is one run.


A six-panel storyboard for a 20-second sequence of an animated feature: numbered pencil panels with the same two characters, lantern and screen direction throughout, each captioned with shot size, action and duration. For directors, storyboard artists and producers who need to agree on a sequence before production.

:::note As of October 2026
The models, settings and screens in this article are the ones Scenario offered in October 2026. New models and updates arrive often, so check the app for newer options before you choose.
:::

![Six pencil storyboard panels in two columns, each with a number, a shot type, an action line and a duration: a workshop, stone steps to a punt, a lantern set on the water, a slipping knot, the lantern drifting toward a bridge, and the girl holding it under the bridge](https://ufgc0c9mhl4ueq76.public.blob.vercel-storage.com/academy/1fd32acf306941f904ba78c45a87d1f344c7da816a320ec7ee5e533eb8c969ff.webp "The finished board. Panels by GPT Image 2.5 Sunburst, captions set in code.")

## The brief

| Field | Brief |
|---|---|
| Goal | A six-panel storyboard for a 20-second sequence, to agree on shots, order and timing before production |
| Script | 1 workshop, Isa shows Master Oren her finished lantern. 2 they go down the steps to the punt. 3 Isa sets the lit lantern on the water on a cord. 4 insert, the knot slips. 5 over the shoulder, the lantern drifts toward the Great Bridge. 6 dusk, under the bridge, Isa holds it up |
| Format | 16:9 panels, loose graphite pencil, captions outside the drawing |
| Fixed | The two characters, the lantern (round globe, exactly three brass bands, ring handle), the canal flowing left to right and the Great Bridge on the right in exterior shots |
| May vary | Pencil texture, background detail, exact poses |
| Accept when | Every panel reads at thumbnail size; cast and lantern match the sheets; screen direction holds; each shot shows who the shot line names, and no one else; no stray text in the art |

Isa, a 12-year-old glassblower's apprentice, and Master Oren live in Quillfern, an invented canal city of glass lanterns. The sheets made here can feed [One scene, many angles](https://learn.scenario.com/playbooks/long-video-from-guided-frames.md).

## Approach

:::flow
Shot list | A language model drafts, you edit | text
Cast sheet | GPT Image 2.5 Sunburst, three views | image
Prop sheet | Sunburst, lantern with exactly three bands | image
Panels | Sunburst, both sheets, shared block | image
Captions and board | Python and Pillow | image
:::

A storyboard fails when a character looks different from panel to panel, or when the camera crosses the line and the canal flows the other way. Both are solved before the first panel: continuity lives in two reference sheets and one shared prompt block, and each panel only adds its shot line.

- **Reference sheets first.** One cast sheet (both characters, three views, a few expressions) and one prop and location sheet, made in color because they serve every later stage. See [Character reference sheet](https://learn.scenario.com/playbooks/character-reference-sheet.md).
- **Give each reference a role.** The shared block says "Image 1 is the cast sheet" and "Image 2 is the prop and location sheet", and what to take from each ([Give references a role](https://learn.scenario.com/browse)).
- **State the continuity rule in words.** "In every exterior shot the canal flows from left to right and the Great Bridge is on the right side of the frame" ([Decide what stays fixed](https://learn.scenario.com/browse)).
- **Name only who is in the shot.** A shared block that describes both characters invites both into every panel. Panel 6 failed exactly that way.
- **Keep text out of the model.** Panel numbers, shot types and durations are set in code ([Let code own what must be exact](https://learn.scenario.com/browse)).

Models, as of October 2026:

- **Capabilities:** text to image, and image to image with two references.
- **Model used:** GPT Image 2.5 Sunburst at `high` quality for the sheets and the panels.
- **Alternative:** GPT Image 2.5 Flare. Compare both on two panels before committing to a whole board ([Compare models on your brief](https://learn.scenario.com/courses/pick-the-right-model)).

## Steps

:::steps
1. **Break the script into shots.** One line per panel: number, shot size, angle, action with directions, duration in seconds (2 to 5). Let a language model draft the list (**Create** > **Text**, or Scenario Genie), then fix it by hand: the model proposes the cuts, you decide them.

   ```prompt
   Break this script into a storyboard shot list. One line per shot: number,
   shot size, camera angle, the action in one sentence, duration in seconds
   (2 to 5). Keep screen direction consistent and say who is on the left and
   who is on the right in every shot. Script: [paste the script]
   ```

2. **Make the cast sheet** ([Lock a character sheet](https://learn.scenario.com/browse)). **Create** > **Images**, **GPT Image 2.5 Sunburst**, **Dimensions** 2048x1152, **Quality** `high`, with Isa's approved portrait in **Reference Images**:

   ```prompt
   Character reference sheet for an animated feature film. The reference
   image is Isa: keep her face, freckles, wavy auburn hair tied back, brass
   goggles on her forehead, cream shirt with rolled sleeves, soot-stained
   brown leather apron, olive trousers and worn brown boots exactly as
   shown. Plain warm light-grey background, no text, no labels.

   Two characters, each shown full body in three views side by side: front,
   three-quarter, back. Same scale, standing on one faint ground line, empty
   hands.

   Left group, ISA: the 12-year-old glassblower's apprentice from the
   reference image.

   Right group, MASTER OREN: a stout elderly glassblower, a head and a half
   taller than Isa, bald with a thick white walrus mustache, round
   leather-rimmed spectacles, rolled shirt sleeves showing strong forearms,
   a long burn-marked dark leather apron over a faded blue shirt, grey
   trousers, heavy boots.

   Bottom row, small: four head close-ups of Isa (proud smile, surprised,
   worried, determined) and two of Oren (warm smile, laughing).

   Style: prestige painterly 3D feature animation, hand-painted textures,
   expressive appealing shapes, soft volumetric light, rich color. Even soft
   studio light. Clean gutters between figures.
   ```

3. **Make the prop and location sheet.** Same settings, with a city mood image as reference. The prompt describes the lantern part by part ("a round amber glass globe the size of a melon ... held by exactly three thin brass bands that run from top to bottom, a small brass ring handle on top, a flat brass base"), then the Great Bridge ("one single wide stone arch"), Oren's workshop (round brick furnace, blue door, stone steps to the water) and the canal with a moored punt. Counting details is what keeps the lantern identical later.

   ![Cast sheet: Isa with auburn hair and brass goggles and Master Oren with a white walrus mustache, each in front, three-quarter and back views, with six head close-ups below](https://ufgc0c9mhl4ueq76.public.blob.vercel-storage.com/academy/869df62ba4dddce5563731063a53e774f36bc6b9a468c7b8bc28181449be3bb7.webp "Cast sheet.")

   ![Prop and location sheet: the amber globe lantern on a table and floating on water, the single stone arch of the Great Bridge hung with lanterns, the workshop with its furnace, and the canal with a moored punt](https://ufgc0c9mhl4ueq76.public.blob.vercel-storage.com/academy/eb2db2532d923b6873ddc7adedfeb5f5afa2431068c462b24b75ad5f4cb05e47.webp "Prop and location sheet.")

4. **Write the shared block.** This text starts every panel prompt, unchanged:

   ```prompt
   Storyboard panel for an animated feature film. Loose graphite pencil
   sketch on plain white paper, quick confident lines, light grey shading
   only, no color, no text, no numbers, no arrows, no panel border. Image 1
   is the cast sheet: Isa, a 12-year-old girl with wavy hair tied back,
   goggles on her forehead, rolled shirt sleeves, a long leather apron,
   trousers and boots; Master Oren, stout and a head and a half taller,
   bald, thick white walrus mustache, round spectacles, rolled sleeves, long
   leather apron. Match their proportions, faces, hair and costume shapes.
   Image 2 is the prop and location sheet: Isa's round glass globe lantern
   with three brass bands and a ring handle on top, the single stone arch of
   the Great Bridge hung with lanterns, Oren's canal-side workshop with its
   round furnace and stone steps, and the canal between tall leaning houses.
   In every exterior shot the canal flows from left to right and the Great
   Bridge is on the right side of the frame.
   ```

5. **Generate each panel** ([Batch from a template](https://learn.scenario.com/browse)). **Reference Images** > **Select from Library**: the cast sheet first, then the prop sheet. **Dimensions** 1536x864. Append one shot line to the shared block, for example:

   ```prompt
   Shot: 5. OVER THE SHOULDER, behind Isa standing in the punt. The lantern
   floats away from her on the current, left to right, toward the Great
   Bridge on the right side of the frame, getting smaller. Isa reaches one
   arm out after it.
   ```

   The six panels ran in parallel, about 45 to 55 seconds each.

6. **Inspect and retake.** Check each panel against the sheets and its shot line. Panel 6 ("Isa stands in the punt holding the lantern high") came back with Oren rowing beside her: he stayed on the steps in panels 3 to 5, and the shot line has Isa alone. The retake removed Oren and the continuity rule from the block for this panel and named the shot's limits:

   ```prompt
   ...Image 1 is the cast sheet: Isa, a 12-year-old girl with wavy hair tied
   back, goggles on her forehead, rolled shirt sleeves, a long leather
   apron, trousers and boots. Match her proportions, face, hair and costume
   shapes. Image 2 is the prop and location sheet: Isa's round glass globe
   lantern with three brass bands and a ring handle on top, and the single
   stone arch of the Great Bridge hung with lanterns.

   Shot: 6. WIDE, dusk, under the arch of the Great Bridge. Isa stands alone
   in the punt holding the lantern high in both hands, its warm light
   filling the underside of the arch and reflecting on the water; the
   hanging lanterns of the parapet glow above. She is UNDER the only Great
   Bridge in the story: no other bridge, arch or viaduct appears anywhere in
   the frame, and the canal beyond the arch opens onto calm water and
   rooftops only.
   ```

7. **Caption and assemble in code** ([Typeset text, never generate it](https://learn.scenario.com/browse)). This run used Python with Pillow: each panel scaled to 1152x648 with a dark outline, a cream strip underneath with the number in Georgia Bold (terracotta) and "shot | action | duration" in Georgia, then two columns on one board. The captions come from the same JSON file as the shot lines:

   ```python
   d.text((24, 703), str(p["n"]), font=georgia_bold_46, fill=(196, 101, 74), anchor="lm")
   d.text((90, 703), f'{p["shot"]}  |  {p["action"]}  |  {p["secs"]} s',
          font=georgia_28, fill=(17, 17, 17), anchor="lm")
   ```
:::

**With the Scenario MCP:** `model_run` on `model_openai-gpt-image-2-5-sunburst` with `referenceImages` set to the two sheet asset ids, `width: 1536`, `height: 864`, `quality: "high"`. Keep the shared block in a file and append each shot line in your script, so an agent cannot paraphrase it between panels.

## Check the result

:::check
**Accept when:** both characters keep their silhouettes (Oren's mustache and spectacles, Isa's goggles and apron); the lantern keeps its globe, bands and ring; the canal flows left to right and the Great Bridge sits on the right; each panel shows exactly who its shot line names; shot sizes differ visibly; the art contains no letters or numbers.
**Reject when:** a character appears where the script does not put them; a landmark is duplicated; the lantern changes shape; a character changes height or costume; the screen direction flips; the model writes captions or numbers into the drawing.
:::

:::compare
![Rejected panel 6: Isa holds the lantern up under the bridge while Oren sits rowing in the punt beside her](https://ufgc0c9mhl4ueq76.public.blob.vercel-storage.com/academy/0d5bdc06732b71765d7902f7d904d59222c6918d11fafb3479f5ae6b3e5fa2d4.webp "Rejected: Oren is in the punt, although he stayed on the steps and the shot line has Isa alone.")
![Kept panel 6: Isa alone in the punt under the arch, holding the lantern up](https://ufgc0c9mhl4ueq76.public.blob.vercel-storage.com/academy/795c4c48f1c997c3e0ecdb3f4238f801c73d6748c5dc824b0f22de7850c8c8c0.webp "Kept: the retake describes only Isa and says the bridge is the only one.")
:::

Three smaller deviations were accepted and written in the board notes:

- **More finished than "loose".** Every panel came back as a detailed pencil rendering. That reads well in a pitch; for a fast rough pass, ask for "thumbnail sketch, under 30 lines".
- **A thick rope in panel 4.** The shot line said "thin cord". The insert still reads as the knot slipping.
- **Distant bridges.** Panels 1, 2, 3 and 5 show smaller bridges far down the canal through the arch, as the location sheet does. The Great Bridge stays the largest and on the right, so they read as the city's other bridges.

:::quiz Check yourself
? Two takes of panel 6. The shot line: "Isa stands in the punt holding the lantern high." Oren stayed on the workshop steps in panels 3 to 5. Which take goes on the board?
![Pencil storyboard panel under a stone bridge: a girl holds a lantern up in a punt while a stout bald man with a mustache sits at the oars](https://ufgc0c9mhl4ueq76.public.blob.vercel-storage.com/academy/0d5bdc06732b71765d7902f7d904d59222c6918d11fafb3479f5ae6b3e5fa2d4.webp "Take A")
![Pencil storyboard panel under a stone bridge: a girl stands in a punt holding a lantern up, the canal and rooftops beyond the arch](https://ufgc0c9mhl4ueq76.public.blob.vercel-storage.com/academy/795c4c48f1c997c3e0ecdb3f4238f801c73d6748c5dc824b0f22de7850c8c8c0.webp "Take B")
- [ ] Take A: the shared block names both characters, so Oren belongs in the shot
- [x] Take B: only the people the shot line names, where the story left them
- [ ] Either one: a storyboard only needs to read at thumbnail size
> A character the script did not put in the shot is a continuity break. The retake removed Oren from the block for this panel and wrote "Isa stands alone".

? Your pencil panels come back with the panel number and shot type written into the drawing. What is the fix?
- [x] Keep "no text, no numbers, no arrows" in the shared block and set every caption in code
- [ ] Add the exact caption text to each shot line so the model spells it right
- [ ] Accept the panels when the numbers are correct and legible
> The art must contain no letters or numbers. Captions set in code are always correct and legible.
:::

## Make it repeatable

- **Save the kit.** Put the two sheets, the shared block and the shot list in one collection. A new sequence with the same cast starts from the same kit ([Organize your work](https://learn.scenario.com/browse)).
- **Build it as a workflow** ([Build your first workflow](https://learn.scenario.com/workflows/build-your-first-workflow.md)). In the **Node Editor** (**New Workflow**):
  - a **Text** node for the shot list, and **Split Text** to turn it into one line per panel;
  - a **Prompt Builder** that joins the shared block and the line;
  - an **Image Generator** on GPT Image 2.5 Sunburst with both sheets wired to **Reference Images**, inside a **Loop** over the lines;
  - an **Approval** node after the generator, so a reviewer can reject a panel before the board is assembled.
- **Keep the captions in code.** The caption script reads the same shot list, so captions always match the prompts.
- **Next stages.** The approved board feeds [One scene, many angles](https://learn.scenario.com/playbooks/long-video-from-guided-frames.md) and [Long video from guided frames](https://learn.scenario.com/playbooks/long-video-from-guided-frames.md).

## Runs and time

| Step | Model | Runs | Kept | Wait |
|---|---|---|---|---|
| Cast sheet | GPT Image 2.5 Sunburst | 1 | 1 | about 45 s |
| Prop and location sheet | GPT Image 2.5 Sunburst | 1 | 1 | about 50 s |
| Six panels, in parallel | GPT Image 2.5 Sunburst | 6 | 5 | about 55 s |
| Panel 6 retake | GPT Image 2.5 Sunburst | 1 | 1 | about 45 s |
| Captions and board | Python and Pillow | 0 | n/a | seconds |

Nine runs at about the same price each, so the board costs the number of panels plus two sheets and any retakes. The price shows on the **Generate** button before you run.

## Variations

- **Brand and marketing:** a 30-second ad storyboard with a product sheet in place of the prop sheet; add the product's fixed details (label side, color) to the continuity rule.
- **Learning:** a training-video board where each panel is one step of a procedure, with the step text set in code under the panel.
- **Games:** a cutscene board; reuse the same shared block to generate the cinematic's key frames later.
- **Different look:** swap "graphite pencil" for "marker rough" or "gray-tone digital thumbnail" in the shared block; the sheets stay the same.
- **Animatic:** time each panel to its duration and add a scratch voice ([Narration and speech](https://learn.scenario.com/audio/narration-and-speech.md)).

## If it breaks

- **A character appears in a shot they are not in.** Describe only the characters the shot names in that panel's block, and write "alone" when one character is alone.
- **A landmark appears twice.** Say there is only one ("the only Great Bridge in the story") and what lies beyond it.
- **Characters drift in pencil.** The color sheet carries cues a pencil panel cannot show. Repeat the shape details in words (goggles, walrus mustache, round spectacles).
- **Text or numbers appear in the drawing.** Keep "no text, no numbers, no arrows" in the shared block and set every caption in code.
- **All panels look like the same medium shot.** Name the shot size, the angle and a camera height in each line; vary them deliberately.

Source: https://learn.scenario.com/playbooks/storyboard-from-script