← Back to all browser tools
FLUX 3 IMAGE / LAYOUT TOOL

Draw the layout.
Copy the FLUX 3 prompt.

Give every element a place. Draw bounding boxes, describe what goes inside, and copy a prompt with the right coordinates for FLUX 3 Image.

Free layout planning for new images. Works with the documented FLUX 3 Image and Klein element-table format. Create the image in your chosen FLUX app after exporting.

Your composition

Drag empty space to draw. Drag a box to move it. Select a box to edit its edges below; arrow keys move a focused box by 1 unit, or 10 with Shift.

Use 4:5 in your FLUX app too.

Elements

0 elements

No boxes yet. Choose Add box to enter coordinates without drawing.

Each box ID becomes a <tag> in the caption. Missing tags are added with their element descriptions. Edit the caption for natural, detailed wording.

Your layout prompt

Caption + element JSON
Element JSON only

How to make a FLUX 3 layout prompt

  1. Choose the frame. Pick the same aspect ratio you will use for generation.
  2. Place your elements. Draw boxes or use Add box and type their edges. Overlapping boxes are allowed.
  3. Describe the image. Give each element its appearance and write a scene caption. Put exact words for a heading in quotation marks.
  4. Copy and generate. Paste the full output into a FLUX workflow that accepts layout prompts. The downloaded JSON contains prompt and aspect_ratio.

FLUX 3 bounding box coordinates

The normalized order is [top, left, bottom, right], with integers from 0 to 1000. Both axes scale independently: [0, 0, 1000, 500] covers the left half of a square or a portrait frame. The full frame is [0, 0, 1000, 1000].

For example, edges at left 384, top 108, right 1536, bottom 972 in a 1920 × 1080 image become [100, 200, 900, 800]. This tool handles the normalized coordinates for you.

Using the layout with FLUX 3 Image Klein

Klein documents the same generation layout format. A self-hosted workflow needs a detailed scene caption: follow the model card’s prompt-upsample instructions, and preserve the element tags and table. This planner assembles your descriptions; it does not run a caption model.

Boxes guide composition; objects can extend beyond them, text can be wrong, and small elements may be missed. The preview shows the planned regions, not a generated image. Existing-image edits use different source and target fields; use the official editing guide for that workflow.

Format references, checked 17 October 2026: BFL bounding box documentation, layout prompting guide, and official Klein model card. Tool Pantry is an independent layout helper.

After generating your image

Prepare the finished file with the image resizer, JPG to WebP converter, or 1 MB image compressor. Keep the original and inspect fine text after resizing or compression.

QUICK ANSWERS
What order do FLUX 3 bounding box coordinates use?

Use [top, left, bottom, right], with whole numbers from 0 to 1000. The vertical coordinate comes first. [0, 0, 500, 500] is the top-left quarter at any aspect ratio.

Can I use this with FLUX 3 Image Klein?

The official Klein model card documents the same caption plus element-table format. Use a FLUX interface or workflow that accepts that prompt format. Self-hosted Klein needs a detailed scene caption; follow its prompt-upsample guidance.

Does this tool generate an image or charge API credits?

It creates a layout prompt and JSON in your browser. Image generation happens separately in your chosen FLUX app or workflow, which may charge for generation.

Are my descriptions saved or uploaded?

The layout and descriptions stay in this tab. This tool does not upload them or save them between visits. Download the JSON before leaving if you want to keep a copy.