Generate images
Work can generate and edit images directly in the chat interface. Describe what you need in plain language and get visuals you can use for presentations, product mockups, marketing materials, and more.
Work uses Black Forest Labs models for image generation.
Activation
- Click the
+icon or type/in the chat window. - Select
Toolsthen enableImage generation.
Image generation stays active for the current session. You can enable other tools at the same time.
Daily generation limits apply based on your plan. For details, visit our pricing page.
Creating images
Describe the image you want in natural language. You don't need complex prompts or special syntax. Say what you're looking for and Work handles the rest.
For the best results, include:
- Subject: what the image shows.
- Style: illustration, photorealistic, diagram, flat design, etc.
- Constraints: aspect ratio / orientation.
Example prompt: "Create a flat illustration of a modern office interior with natural light and plants."
Editing images
You can refine generated images or edit existing ones in two ways.
Continue in the same chat
After generating an image, describe the changes you want in a follow-up message. Work understands you're referring to the previous image and produces an updated version.
Example: "Add a whiteboard on the back wall. Keep the lighting and color palette unchanged."
Upload an image in a new chat
Start a new chat, then upload an image using the + button by selecting Upload Files.
Describe the edits you want, and Work generates a new version based on your instructions.
This is useful when you want to pick up a previous result or edit an image from another source.
Prompt templates
Here are reusable patterns for common use cases:
| Type | Template |
|---|---|
| Scene | Subject, style (flat/minimal), composition, aspect ratio |
| Photographic | Subject, photorealistic style, lighting type, lens, aspect ratio |
| Product mockup | Subject, studio shot, background, composition, aspect ratio |
| Diagram | Subject, clean lines, readable labels, accessible colors, aspect ratio |
| Presentation visual | Subject, slide-friendly composition, brand colors, minimal text, aspect ratio |
| Marketing material | Subject, target audience tone, style, brand guidelines, aspect ratio |
Common use cases
- Pitch and deck visuals: hero images, section dividers, and stylized backgrounds for investor decks or internal presentations.
- Marketing assets: campaign banners, landing-page heroes, social posts with consistent brand tone.
- Product mockups: render a concept on a neutral background to compare directions before a designer takes over.
- Internal communications: header imagery for announcements, onboarding decks, all-hands recap docs.
- Concept exploration: generate multiple visual directions from a brief before committing one to production.
- Diagrams and conceptual illustrations: process flows, organizational charts, abstract concept visuals for documents and reports.
- Placeholder visuals during drafting: get a stand-in image immediately while a final asset is in production.
Best practices
- Be specific. "A photorealistic hero image for a fintech landing page, dark background, abstract light trails" produces better results than "make a banner."
- Iterate. Treat the first result as a starting point. Refine with follow-up prompts until you're satisfied.
- Pair with professional tools. Image generation works best alongside dedicated editing software for final adjustments like precise cropping, color correction, or brand overlays.
- Download and reuse. Click the download button on any generated image to save it locally.
For charts, graphs, and data visualizations, use Code Interpreter in Chat instead. It generates accurate visuals from your actual data using Python.
Limitations
- Prompts or images that contain copyrighted or inappropriate content may be blocked by safety filters.
- Fine details (small text, complex patterns) can sometimes be lost or altered during editing.