How to use AI Image Generation?

Chatly’s AI Image Generation feature lets you turn ideas, descriptions, or photos into stunning visuals instantly.

You can:

• Start with a text prompt or reference image.

• Choose from multiple models optimised for generation and editing.

• Adjust aspect ratios to fit your format (ads, print, stories, etc.).

• Refine and edit results directly using prompt adjustments or model variations.

Step-by-Step guide to AI Image generation

Step 1. Start a Chat

• Open a chat with the Image Agent

• Enter your prompt (e.g., “a modern office with natural light”)

• Optionally attach up to 2 reference images

Step 2. Image Agent Generates Results

• The agent processes your input (prompt + references)

• You’ll receive 2 generated images per request

Step 3. Edit or Refine

• In conversational view, you can request edits on the last 2 generated images.

• In full view, you can focus on a single image and edit that specific one.

• You can also upload your own image to edit.

Step 4. Continue the Conversation (Image agent)

• Ask follow-up questions (e.g., “make the lighting warmer” or “add plants”)

• The image agent may ask clarifying questions if your prompt is unclear

Step 5. Edit Image

• Preview results in full view.

• You can choose to apply the following enhancements to the image:

  • Upscale: Increase the resolution of the image and add more richness.

  • Background Remove: Choose to remove the background from the generated image.

  • Edit Image: Chat to edit the image.

Note

• For enhancements such as Upscale and Background Removal, tokens are deducted separately from standard image generation. Each enhancement is counted as a new request with its own token cost.

Step 6. View & Download

• Preview results in full view.

• Download the preferred generated image.

Step 7. History

• View all generated images in your history.

• Continue chat from the same point with context and generate new images.

Tips for Best Results

• Be descriptive: Instead of “a cat,” try “a fluffy orange cat sitting by a sunny window.”

• Use aspect ratios: Some models support defining width × height for more control (e.g., “16:9” for landscapes).

• Leverage references: Attach up to 2 reference images to influence style, composition, or subject.

• Iterate with edits: Start broad, then refine (e.g., generate a cityscape → edit to add neon lights → refine to make it night-time).

• Model choice matters: Different models have strengths. Try switching if you want a different style or detail level.

• Keep prompts natural: Write in plain language — the system is designed to understand conversational instructions.

Image Generation Limits by Plan & Duration

Plan	Billing Cycle	Tokens	Images	Generations
Standard	Monthly	1,100	~47	~23
	Quarterly	2,200	~95	~47
	Yearly	4,250	~184	~92
Professional	Monthly	4,250	~184	~92
	Quarterly	10,000	~434	~217
	Yearly	20,000	~869	~434

You get 100 tokens FREE on signup for image generation. 🎉

(Tokens refresh with your plan’s billing cycle monthly, quarterly, or yearly.)

FAQs

Q: What’s the difference between Conversational view and full view?

• Conversational View: Conversational mode where you see the last two generated images in context with the chat.

• Full View: Focused mode where you preview a single image in isolation and can edit that specific one.

Q: Can I edit my images?

Yes! You can edit your images by attaching them in the prompt box as a reference. You can edit the last two images generated in the chat. If you’d like to edit an older image, simply open it in full view or download it first, then reattach it as an input.

Q: How are tokens counted?

• Tokens are based on the model used, and apply equally to generation and editing

• Each request = 2 images, so total tokens = per-image cost × 2