PROMPT STUDIOImage workspace ↗Account ↗

Opening your studio…

Checking your account.

ART & PROMPT FIELD GUIDE

Know what you are mixing.

Movement weights express art direction, not hidden model settings. Combine deliberately, then review the results.

Edit the visual system

Choose a visual reference

Attach the same reference when using prompts in different models.

Analysis calls the vision model. You will review the result before replacing your current document.

Saved versions

Restore as a draft. Saving it creates a new revision.

MODEL HANDOFF

Build a consistent collection.

Batch variables · import a collection

CSV columns: subject (required), occasion, message, audience. One artwork per row, up to 100 rows. Prompt expansion does not submit images.

Use the same reference for the closest match. Different models can still vary; check color, texture, composition and text before approving a series.

The image workspace runs xAI. GPT Image and Gemini prompts are exported for use in those tools. Continuing creates a project; images are generated only when you click Generate there.

Developer handoff

Export a working package your team can validate and keep alongside its code.

This package uses the Odd Palette format. It is not a native Photoshop, Figma or ComfyUI file. Unsaved edits are exported as a draft.

VISUAL CONSISTENCY LAB

One style. Three different subjects.

Test a saved revision against real images before using it for a larger series.

Generate one image at a time. Each Standard image costs 2 credits; 3 images cost 6 credits. Your balance is checked before each image.

Creating a set does not generate or charge for images.

Previous test sets

Exact test prompt

Use this prompt with the same visual reference. Keep the target tool’s actual model name and settings with your results.