- ArtCraft text to image turns natural-language prompts into refined images with editing controls.
- Prompt structure matters: subject, environment, composition, lighting, then style.
- Reference images guide pose, layout, and appearance for consistent results.
- Iterate one variable at a time instead of rewriting the whole prompt.
- Access via the browser workspace or desktop app.
How ArtCraft Text to Image Works
ArtCraft is an open-source AI creative studio for artists, designers, and filmmakers. Its text to image workflow combines natural-language prompting with reference-driven generation and in-place editing, so you build an image progressively rather than gambling on a single prompt. You can work in the web workspace or through the desktop application.
The core capability set includes four pillars:
| Capability | What It Does | Best For |
|---|---|---|
| Text-to-image | Generates new images from a written prompt | Original concepts, style exploration |
| Reference guidance | Uses an imported image as visual direction | Poses, identity, layout consistency |
| Image editing | Changes selected elements, keeps the rest | Fixing details, controlled variations |
| Background replacement | Swaps the environment around the subject | Product and portrait alternatives |
Write a short, focused prompt first. Add composition, lighting, and style details only after you have a result worth refining.
Prompt Structure That Actually Works
Effective prompting in ArtCraft is about organizing creative intent, not stacking keywords. Use this structure as your foundation:
[subject], [appearance or action], in [environment], [composition and viewpoint], [lighting and mood], [visual style]
The table below breaks down each slot:
| Slot | Purpose | Example Fragment |
|---|---|---|
| Subject | One primary focus | "a lone explorer" |
| Appearance / action | What the subject looks like or does | "wearing a weathered coat, looking back" |
| Environment | Where the scene happens | "in a foggy pine forest at dawn" |
| Composition | Framing, viewpoint, camera | "wide shot, low angle, subject off-center" |
| Lighting and mood | Atmosphere and tone | "soft rim light, cold mist, quiet mood" |
| Style | Visual language | "cinematic, muted palette, film grain" |
Text-to-Image
- Full creative freedom
- Best for new concepts
- Iterate through variations
Reference-Led
- Visual anchor from an image
- Control pose and layout
- State what to keep vs. change
Edit and Refine
- Targeted changes
- Preserve the broader concept
- Great for detail corrections
Changing every part of the prompt at once makes results impossible to compare. Keep successful elements, refine one issue, and generate again.
Step-by-Step: Your First Generation
Open the Workspace
Access ArtCraft through the web experience or desktop app and enter the main creative workspace where projects and generation tools live.
Choose an Image Workflow
Select an image-oriented workflow and a starting canvas. Decide whether you begin from text alone or from an imported reference.
Write the Prompt and Add References
Describe the subject, environment, composition, and lighting. Optionally import a reference image and state which elements to preserve or change.
Generate and Review
Run the generation, inspect the output, and compare it against your creative direction. Adjust one variable at a time.
Edit and Export
Use ArtCraft's editing tools for final adjustments, then save or export the finished image.
Identify the strongest parts of your result, choose a single issue to improve, preserve composition and subject details, and compare versions side by side.
Beyond Flat Images: Editing and 3D Composition
Text to image is the entry point, but ArtCraft's power comes from combining it with editing and spatial workflows.
| Workflow | Approach | Best For |
|---|---|---|
| 2D compositing | Arrange and adjust flat elements | Posters, layered illustrations, social graphics |
| Background processing | Separate or replace backdrops | Cutouts, focused portraits, product shots |
| 3D compositing | Place subjects in spatial scenes with camera control | Scene mockups, depth-rich compositions |
| Scene control | Manage positioning and visual relationships | Repeatable layouts, precise staging |
For 3D work, the workflow moves from importing an image or preset, to arranging foreground, middle ground, and background, to setting the camera angle and distance, and finally generating the spatial composition. Structure those prompts around depth relationships: [subject] positioned in [location], with [foreground elements], [background elements], viewed from [camera angle and distance].
Development discussion and roadmaps happen on the official Discord and the GitHub repository.
Pre-Generation Checklist
Before You Generate:
- Define one clear primary subject
- Describe the environment and context
- Specify framing, viewpoint, and lighting
- Prepare reference images if visual guidance is needed
- Plan to refine one variable per iteration
FAQ
Q: What should I check when an ArtCraft text to image generation fails?
Review the prompt and selected settings, confirm required inputs are available, and try a simpler generation request. If the issue continues, reload the application and retry before contacting support.
Q: Do I need a reference image for text to image generation?
No. References are optional. They help when you need visual consistency for pose, layout, or appearance, but pure text prompts work well for exploring new concepts.
Q: How is editing different from regenerating?
Editing changes selected visual elements while keeping the rest of the image intact, which is ideal for detail fixes. Regenerating produces a new result from the prompt, better for exploring alternatives.
Q: Where do I get account or product support?
Use the official ArtCraft support page at getartcraft.com/support and include a concise problem description, the workflow you were using, and any error details.