Generate Images
Your employees can create images from text descriptions. Logos, diagrams, social media graphics, illustrations, and more.
TL;DR
Every employee can generate images from text descriptions. Describe what you want and the result is saved to Drive and shown inline in chat. No setup required. You can set the default engine, quality, and shape once in the tool settings, or let the employee choose per image.
How It Works
You describe the image in natural language. The employee enhances your prompt, generates the image, saves it to your Drive, and shows it inline in the chat.
| Step | What Happens |
|---|---|
| 1. You describe | "Create a minimalist logo for a coffee brand called Brewline" |
| 2. Employee refines | Enhances your prompt with details for better results |
| 3. Image generates | Image is created via an AI image generation model |
| 4. Saved to Drive | Persistent URL stored in the employee's Drive tab |
| 5. Shown in chat | Image appears inline. Click to view full size |
What It Can Do
| Capability | Example |
|---|---|
| Marketing visuals | "Create a social media banner for our product launch" |
| Logos & branding | "Design a minimalist logo for a tech startup" |
| Illustrations | "Draw an isometric office scene with people working" |
| Diagrams | "Create a visual showing our 3-step onboarding process" |
| Mockups | "Generate a hero image for a SaaS landing page" |
How to Set It Up
Nothing to do. Image Generation is enabled for every employee at hire.
You can ask the employee to manage their own tools, or do it manually:
- Select the employee
- Click the Tools tab
- Find "Image Generator" in the Actions section
| Action | What it does | How |
|---|---|---|
| Enable / Disable | Controls whether the employee can use this tool | Toggle the switch |
| Defaults | Set the engine, quality, and shape the employee uses by default | Open the tool and pick from the Defaults panel |
| Tool Rules | Custom instructions that guide how the employee uses this specific tool, e.g. "always use flat design" or "stick to our brand colors: navy and gold" | Expand the tool, then write your rules in the text field |
| Delete | Permanently removes the tool from the employee | Click the delete button |
Set Your Defaults
Open the Image Generator and you control the baseline the employee starts from. Each option shows its credit cost, so you know what you pay before you pick. The employee keeps your defaults and only changes one when a request calls for it.
| Default | Choices | Notes |
|---|---|---|
| Model | A few engines, each with its per-image cost | The employee tells you which one it uses when you ask |
| Quality | Low, Medium, or High | Higher is sharper and costs more. High is the default |
| Shape | Square, Landscape, or Portrait | Match it to where the image goes |
Tips & Tricks
- Be specific. "A flat-design illustration of a woman using a laptop in a modern office, blue and white color scheme" beats "woman at computer"
- Mention style. "Minimalist", "isometric", "watercolor", "photorealistic", "flat design"
- Specify colors. "Use our brand colors: navy blue and gold"
- Include context. "For a SaaS landing page hero section" helps the employee pick the right tone
- Iterate. "Make it more vibrant" or "Try a darker background". The employee refines based on feedback
Behind the Scene
| Powered by | OpenAI image models (GPT Image and DALL-E) |
| How it works | The employee enhances your prompt for better results, sends it to DALL-E, then saves the image to your Drive |
Size Options
You can request a specific size or the employee picks one based on context:
| Size | Best For |
|---|---|
| Square | Social media posts, profile images, icons |
| Portrait | Stories, mobile screens, posters |
| Landscape | Blog headers, banners, presentations |
What It Costs
| Cost | Runtime credits based on processing time |
| HD quality | Costs slightly more credits |
Is It Safe
- Content moderation. The image model enforces content policies automatically. Requests for violent, explicit, hateful, or misleading imagery are blocked before generation
- No personal likenesses. The model will not generate images of real, identifiable people
- No credentials or sensitive data. Image prompts are sent to a third-party AI API. Don't include passwords, API keys, or confidential information in your descriptions
Good to Know
- Content safety. The employee will not generate images containing nudity, violence, hateful symbols, or disturbing content. These guardrails are enforced automatically
- Not a photo editor. Image generation creates new images from scratch. It can't edit existing photos or remove objects from images
- One image per request. Each generation creates a single image. Ask multiple times for variations
- Saved automatically. Every generated image is stored in the employee's Drive tab. You can download it anytime
- Prompt enhancement. The employee may add detail to your description to get better results. This is intentional. They're optimizing the prompt for better output
Frequently Asked Questions
Q: Can the employee edit or modify an existing image? A: No. Image generation creates new images from text descriptions. It can't modify, crop, or edit existing photos.
Q: Where are generated images stored? A: In the employee's Drive tab. Every image gets a persistent URL you can share or download.
Q: Can I specify exact brand colors or fonts? A: You can describe colors ("use navy blue and gold") and the model will approximate them. Exact hex codes and fonts aren't supported. For pixel-perfect brand assets, use a design tool.
Q: How much does image generation cost? A: Each generation uses standard credits. HD quality and larger sizes cost slightly more. Check the Credits page for exact rates.
Q: Can the employee generate multiple variations? A: Yes. Ask "Give me 3 variations of..." and the employee will generate each one separately. Each counts as a separate generation.