# Stability AI MCP

> Stability AI gives your agent access to professional image generation tools. You can create photorealistic images from text prompts, upscale low-resolution graphics while keeping detail, or professionally remove backgrounds from product photos. This MCP lets you handle the entire visual media pipeline right through conversation.

## Overview
- **Category:** ai-frontier
- **Price:** Free
- **Tags:** generative-art, image-generation, upscaling, background-removal, visual-media, computer-vision

## Description

This connector brings industry-leading generative visual capabilities into your existing LLM workflows. Instead of jumping between multiple design tools and juggling complex APIs, you simply talk to your agent. You can ask it to generate a full mockup for an ad campaign from scratch using only text prompts. Need better quality? Just tell it to upscale the image later. Or maybe you've got product photos that need cleaning up; this MCP handles background removal instantly. It’s all about iteration: generating drafts, refining details, and improving resolution—all in one place. If your team uses Vinkius for other services, adding Stability AI gives you a single point of access to professional visual creation.

## Tools

### generate_core_v2
Creates an image using the fast and quality-optimized Stable Image Core model.

### generate_sd35
Generates images using Stable Diffusion 3.5, allowing you to choose between three specific models (large, turbo, or medium).

### generate_ultra_v2
Produces high-end photorealistic images best suited for final production assets.

### get_credit_balance
Checks your current usage quota and available Stable AI credits.

### image_to_image_v1
Modifies an existing picture by transforming it based on a new text prompt.

### inpaint_image
Changes specific, small regions within an image using a prompt while keeping the rest of the photo intact.

### list_engines
Lists all available engine IDs you can use for generating images with v1 tools.

### remove_background
Automatically strips the background from any given photograph.

### text_to_image_v1
Generates an image from scratch using a text prompt and specified size parameters via v1 engines.

### upscale_image
Increases the resolution of an image while mathematically preserving its visual details.

## Prompt Examples

**Prompt:** 
```
Generate a wide format concept visual depicting a sleek, futuristic electric bike stationed alongside a minimalist architectural wall structure.
```

**Response:** 
```
Formatting inputs processing core parameters executing the request systematically through the established model engine...

- **Processing Protocol**: Core Generation (`generate_image`)
- **Render Resultant Instance**: [Image Received via API Data Successfully Base64]

Do you wish to iterate computationally adjusting resolution density with `upscale_image`, or are requirements sufficiently fulfilled?
```

**Prompt:** 
```
Upscale this low-resolution image of a landscape without losing structural fidelity.
```

**Response:** 
```
Command processed. Executing `upscale_image` logic... Done. Resolution enhanced successfully.
```

**Prompt:** 
```
Remove the background from this product photography.
```

**Response:** 
```
Executing `remove_background` routine... Done. Background eradicated cleanly.
```

## Capabilities

### Generate Images from Text
Create entirely new images using advanced models based on detailed written instructions.

### Improve Image Resolution
Increase the size and detail of existing graphics without losing structural quality or fidelity.

### Remove Backgrounds
Automatically clean product photos by isolating the subject and removing its background.

### Edit Specific Image Areas
Replace isolated parts of a photo, like blemishes or missing elements, while keeping the surrounding area consistent.

## Use Cases

### Updating a product catalog in bulk
An e-commerce manager needs 50 shirts photographed against different backdrops. Instead of hiring a photo studio, they instruct their agent to run the `remove_background` tool on all photos, then ask the agent to generate new background images and composite them automatically.

### Creating rapid ad campaigns
A digital marketing specialist needs 20 different futuristic car mockups for A/B testing. They use `generate_sd35` with a specific prompt, asking the agent to cycle through multiple models and generate variations quickly.

### Reviving old artwork
A creative director has an old concept sketch that is too low-res. They use `upscale_image` on the sketch, then refine specific details using `inpaint_image` to turn it into a high-resolution pitch piece.

### Developing complex mockups
A product designer needs an image of a smart speaker in a modern living room. They prompt the agent to generate the initial mockup using `generate_core_v2`, and then follow up by asking it to adjust the lighting or remove extraneous objects.

## Benefits

- You stop worrying about which generation model to use. You just ask for the result you want—be it `generate_core_v2` speed or `generate_ultra_v2` quality—and your agent handles the selection.
- Need to adapt an existing mockup? Use `image_to_image_v1` to change the style of a generated image, keeping its composition while applying a new look. It’s powerful iteration in one prompt.
- Product photography just got easier. The `remove_background` tool cleans up product shots instantly, standardizing your catalog images without Photoshop.
- Low-resolution assets don't have to be discarded. Simply run `upscale_image` and get a high-detail version that can actually be used in print or digital ads.
- The system handles complex edits with tools like `inpaint_image`, letting you fix small errors or add missing details to an image without having to mask off the area first.

## How It Works

The bottom line is that instead of running three different processes in separate software, your AI client manages the entire creative workflow for you.

1. Connect your API key to this MCP within your agent's settings.
2. Tell your agent exactly what you need. For example: 'Generate a futuristic bike mockup and then remove its background.'
3. Your agent executes the necessary steps using multiple tools, delivering the completed, refined visual asset.

## Frequently Asked Questions

**How do I use Stability AI with my existing images?**
You can transform or enhance old photos using `image_to_image_v1`. This tool lets you apply a new style to an image while keeping the original structure intact.

**Is the generate_ultra_v2 model best for everything?**
No. While `generate_ultra_v2` is great for final assets, if speed is your main concern, you might prefer using models like those in the `generate_core_v2` tool.

**Can I generate images at a specific size?**
Yes. When generating from text prompts via v1 engines, you can provide explicit width and height parameters to control the output dimensions.

**What if my product image has complex elements in the background that need fixing?**
Use `inpaint_image`. This tool lets your agent surgically target a specific area of the photo, allowing you to clean up or replace objects without affecting the surrounding detail.

**Do I have to pay for every generation with Stability AI MCP?**
No. You can check your current consumption and remaining budget by calling the `get_credit_balance` tool before starting a large batch of operations.