• NEW
  • REFINED
  • FIXED
  • HUMAN

How to generate and edit visual assets with your AI

Connect Stability AI to your agent to move from a text prompt to a finished, polished image. You can build new visuals from scratch or fix specific parts of an existing one.

Ask AI about this page

Short answer

How can I make and fix images with this AI?

Your agent uses text-to-image and image-to-image tools to create high-resolution visuals from your descriptions. Once an image is made, you can use inpainting to edit specific areas without changing the whole picture. This breakdown shows how to handle the full creative loop.

The creative outcomes

Where your ideas land.

Depending on what you ask, your agent will produce one of these results.

NEW

Fresh image generation

The AI builds a brand new visual from a text prompt. You get a high-resolution file that matches your description.

REFINED

Style and structure shifts

Use an existing image as a starting point to change its composition or style. The AI keeps the core idea while applying your new instructions.

FIXED

Localized area editing

Target a specific part of an image to add, remove, or change an object. The rest of the image stays exactly as it was.

HUMAN

Final creative direction

The AI handles the heavy lifting of rendering pixels. You decide which prompts work and which edits are necessary to get the look right.

The workflow

What your AI does when the request arrives.

The agent follows a logical path from a raw idea to a finished asset.

  1. Initial creation

    The agent takes your text description and renders a high-resolution image from nothing.

    generate_sd35generate_ultra_v2

  2. Visual translation

    If you provide a reference image, the agent uses it to guide the new generation.

    image_to_image_v1

  3. Targeted fixes

    The agent modifies specific segments of an image to correct errors or add details.

    inpaint_image

  4. Basic drafting

    For simpler requests, the agent uses standard text-to-image logic to get a quick result.

    text_to_image_v1

Try it

Copy these to start.

Use these prompts to kick off the process with your agent.

Starting points

Swap out the bracketed descriptions for the specific subjects or styles you want to see.

Stability AI Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_eWY2wqnSzJZPVvVlS9FdYkEOD7e86Z3L5fYYGmDm/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — Stability AI capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "stability-ai-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_eWY2wqnSzJZPVvVlS9FdYkEOD7e86Z3L5fYYGmDm/mcp"
    }
  }
}
Copy into chat04

  • Generate a high-resolution image of [a futuristic cyberpunk city street] using generate_ultra_v2.

  • Take this uploaded image and use image_to_image_v1 to make it look like [a watercolor painting].

  • Use inpaint_image to add [a small red bird] to the tree in this photo.

  • Create a simple sketch of [a minimalist coffee cup] using text_to_image_v1.

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect Stability AI once, then ask.

You only need to link your account once. Your credentials stay encrypted, and you can immediately start prompting your agent to create visuals.

Connect Stability AI to your AI

FAQ

How the task behaves

  • 01

    Can the AI edit text inside an image?

    No. The tools are designed for visual elements and objects, not for rendering specific, legible typography.

  • 02

    Can I ask the AI to upscale an existing photo?

    No. The available tools focus on generation and inpainting, not standalone resolution upscaling.

  • 03

    Does the AI remember my previous images?

    The AI can only work with images you provide in the current chat context or those it just generated.

  • 04

    Can the AI create videos?

    No. This Connector is strictly for generating and editing still visual assets.

  • 05

    Who decides if an image looks good?

    You do. The AI follows your prompts, but you are responsible for reviewing the output and providing further instructions for edits.

  • More questions about Stability AI? See everything the Stability AI Connector can do

Connect Stability AI to Claude, Cursor, ChatGPT & more