Generate Similar

Generation Mode Choose how the image should start.

Reference Image Required in image-to-image mode

No file selected

Supported: JPG, PNG, WEBP. Image-to-image adds 1 extra credit.

Prompt

2x2 grid, 16:9, do this for 4 famous movies: -- input: $ subject

create temp view subject_analysis as
select
  infer_main_identity(:subject)          as identity,
  infer_dynamic_pose(:subject)           as action_pose,
  infer_power_or_function(:subject)      as motion_system,
  infer_world_context(:subject)          as world,
  infer_symbolic_details(:subject)       as symbols,
  infer_study_points(:subject)           as closeups,
  infer_short_title(:subject)            as title;

create temp view parchment_board as
select
  build_action_study_sheet(
    material = 'aged cream paper',
    borders = 'thin technical crop marks and panel lines',
    texture = 'ink stains, paper grain, pencil construction marks'
  ) as board
from subject_analysis;

create temp view main_action as
select
  render_manga_subject(
    identity,
    pose = action_pose,
    linework = 'black brush ink, scratch hatching, broken speed strokes',
    energy = motion_system,
    accent_color = 'muted blue, red, or sepia inferred from subject'
  ) as hero
from subject_analysis;

create temp view atlas_layer as
select
  add_world_fragments(
    world,
    style = 'travel-poster landmarks, tiny city shapes, map labels, symbolic scenery',
    placement = 'behind and around the main action'
  ) as background_world,

add_title_and_stamps(
    title,
    style = 'bold poster heading, small seals, side labels, handwritten notes'
  ) as title_system
from subject_analysis;

create temp view study_layer as
select
  add_inset_panels(
    closeups,
    panels = array[
      'face or focal feature close-up',
      'motion diagram',
      'tool or accessory detail',
      'alternate state',
      'silhouette thumbnail'
    ],
    style = 'manga production notes'
  ) as studies
from subject_analysis;

render
  parchment_board.board,
  atlas_layer.background_world,
  main_action.hero center,
  study_layer.studies around main_action.hero,
  atlas_layer.title_system
with
  camera = 'flat illustrated production-board view',
  palette = 'cream paper, black ink, gray wash, one restrained accent color',
  mood = 'dynamic, analytical, legendary, worldbuilding-rich';

-- output rules:
-- 1. $ subject becomes the hero of a manga-style action dossier.
-- 2. surround it with travel-poster-like world fragments and symbolic location details.
-- 3. include inset panels for anatomy, expression, movement, tools, and alternate states.
-- 4. use callouts, arrows, stamps, and notes as decorative production language.
-- 5. preserve subject identity while making the page feel hand-drawn, kinetic, and collectible.
-- 6. avoid polished photorealism, cute toy clay, full-color comic clutter, and empty layouts.

2691 / 10000

Based on: A manga-style action dossier frames a flying superhero in dynamic pose

Ready when you are

Generating your image

Image ready

We could not finish this image

Generating your image

Image ready

We could not finish this image

Send people to the full prompt and image