Built-in AI Image Generation Tool

Learn about MaiAgent's built-in AI image generation tools and compare their results and use cases

📋 Feature Overview

MaiAgent includes three AI image generation tools. They can create new images or edit existing ones using reference images. Select them in your AI agent's tool settings to start generating images.

On the AI agent settings page, click Select Tools under Tools. Under Built-in Tools, select any of these image generation tools:

• Gemini 3 Pro Image

• Gemini 3.1 Flash Image

• GPT Image 2 image generation tool

Image generation tools in the Select Tools dialog

See Configure tools for an AI agent.

---

🎨 Tool Comparison

Tool	Key features	Best for
Gemini 3 Pro Image	Context understanding, Chinese prompts, reference images, conversational editing	Creative work requiring context understanding or Chinese prompts
Gemini 3.1 Flash Image	Fast generation, Chinese prompts, reference images, multi-step editing instructions	Rapid iteration and brainstorming
GPT Image 2	Professional quality, iterative refinement, reference images, transparent backgrounds	Brand design and professional use

⭐️ Generation Example

GPT Image 2: Professional Design Example

Prompt: "Draw a minimalist coffee cup logo with a transparent background, 1024x1024"

Coffee cup logo generated by GPT Image 2 (transparent background)

Highlights: Transparent background, refined lines, and suitable for brand use

---

📝 How to Use the Tools and Best Practices

Basic Prompt Examples

Use case	Example prompt	Recommended tool
Everyday creation	Draw a cute puppy	Gemini 3 Pro Image
Professional design	Design a modern minimalist logo with a transparent background	GPT Image 2
Rapid prototyping	Quickly generate a concept image for a website home page	Gemini 3.1 Flash Image

Multi-round Refinement (GPT Image 2)

Round 1: "Design a logo for a coffee shop"
Round 2: "Change the color to dark brown"
Round 3: "Add some steam"
Round 4: "Make the overall design more minimalist"

Edit Using a Reference Image

Gemini 3 Pro Image, Gemini 3.1 Flash Image, and GPT Image 2 all support editing with reference images. Upload the image you want to change in the conversation and describe only the changes you want. Unmentioned parts will remain as unchanged as possible:

"Use this image as a reference and change the background to a beach scene"
"Keep the person unchanged and only change the clothing color"
"Add some flowers to this scene"

If the AI agent says it cannot access a previously generated image, upload that image again in the conversation before requesting changes.

---

🎯 Practical Use Cases

Use Case 1: Social Media Content

Need: Create an image for an Instagram post

Recommended: Gemini 3 Pro Image

Example prompt: Create a cozy cafe scene for an IG post

Use Case 2: Corporate Brand Design

Need: Design a company logo and brand assets

Recommended: GPT Image 2

Example prompt: Design a minimalist, modern logo for a technology company with a transparent background

Use Case 3: Brainstorming and Prototyping

Need: Quickly visualize ideas and refine them iteratively

Recommended: Gemini 3.1 Flash Image

Example prompt: Concept art for a mobile app interface design

---

❓ Frequently Asked Questions

Image Quality

Q: How can I get the highest-quality images?

A: Use GPT Image 2 or Gemini 3 Pro Image and provide a detailed description:

• ✅ Specify the style (for example, "professional photography")

• ✅ Describe the scene in detail (lighting, angle, atmosphere)

• ✅ State your quality requirements (for example, "high resolution" or "commercial quality")

Q: Why doesn't the generated image match my expectations?

A: Try improving your prompt:

• 🎯 Use concrete details instead of abstract descriptions

• 🎨 Specify the artistic style

• 📐 Describe the composition and viewing angle

• 🌈 Describe the colors and lighting

Features

Q: How do I generate an image with a transparent background?

A: Use GPT Image 2 and explicitly request a transparent background:

Design a logo with a transparent background
Create an icon with a transparent background

Q: Can I edit an image I have already generated?

A: Yes. Gemini 3 Pro Image, Gemini 3.1 Flash Image, and GPT Image 2 all support reference image editing. Upload the image you want to change, then describe the change:

Use this image as a reference and change the sky to sunset colors
Keep the composition unchanged and only change the person's clothing