TheGlitch

Posted on Mar 5

Image Processing in Your AI Workflow: TheGlitch MCP Integration

#mcp #ai #imageprocessing #webdev

If you've been following MCP (Model Context Protocol) adoption over the past few months, you've probably noticed a pattern: the most useful MCP servers are the ones that replace something you were already doing manually.

TheGlitch just became one of those.

Quick Background

TheGlitch is a stateless image processing API — REST endpoints for resize, format conversion, effects, AI background removal, optimization, and social media presets. Nothing stored, everything processed in memory and returned as binary.

It's been on RapidAPI since February. What's new is that it's now available as an MCP server, which means you can call it from Claude Desktop, Claude Code, Cursor, VS Code, or any MCP-compatible client using natural language.

Setup (90 Seconds)

1. Get a RapidAPI key

Free plan: 500 requests/month, no credit card.
→ rapidapi.com/theglitchapp/api/theglitch-image-processing

2. Add MCP config

{
  "mcpServers": {
    "TheGlitch Image Processing": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://mcp.rapidapi.com",
        "--header",
        "x-api-host: theglitch-image-processing.p.rapidapi.com",
        "--header",
        "x-api-key: YOUR_RAPIDAPI_KEY"
      ]
    }
  }
}

Config file location by client:

Client	File
Claude Desktop	`~/Library/Application Support/Claude/claude_desktop_config.json`
Claude Code	`.mcp.json` (project root)
Cursor	`.cursor/mcp.json`
VS Code	`.vscode/mcp.json`

Node.js required (for npx). No manual package installs — npx fetches mcp-remote automatically on first run.

3. Restart your AI client

That's it.

What's Exposed

Seven endpoints map directly to MCP tools:

Tool	What It Does
`process`	Full pipeline: resize + effects + format conversion in one call
`resize`	Resize to width, height, or both with a resize mode
`convert`	Convert between JPEG, PNG, WebP, GIF, BMP
`effects`	Brightness, contrast, saturation, blur, sharpen, grayscale, sepia, rotate, flip
`remove-bg`	AI background removal
`optimize`	File size reduction with quality control
`preset`	Social media presets (instagram-square, facebook-cover, youtube-thumbnail, etc.)

Prompt Examples

These aren't hypothetical — these are real natural language prompts that map cleanly to API calls:

Resize https://example.com/photo.jpg to 800x600 and convert to WebP

Remove the background from this product image

Apply grayscale and increase contrast by 40 to this image

Convert this image to instagram-square format

Optimize this image for web — target 80% quality

The AI resolves the intent, calls the correct endpoint with the correct parameters, and returns the result.

Practical Workflow Examples

Scenario 1: Documentation screenshots

You have 20 screenshots to prepare for docs. Instead of running them through an external tool:

"Resize all images in /docs/screenshots to 1280px wide and convert to WebP"

The AI iterates, calls the API for each file, handles the responses.

Scenario 2: Product image pipeline

"Take product.png, remove the background, then create three versions: instagram-square, facebook-cover, and a 800x600 web version"

Three sequential MCP calls, one instruction.

Scenario 3: Pre-commit image optimization

"Before I commit, optimize all images in public/assets/ to WebP at 85% quality"

How Authentication Works

MCP calls use the same RapidAPI key as direct REST calls. Same plan, same rate limits, same usage counter. No separate key management.

Under the hood, mcp-remote proxies through mcp.rapidapi.com, which RapidAPI hosts. Your key authenticates both channels — you don't need to set up anything separately.

Available Plans

Plan	Requests/month	Price
BASIC	500	Free
PRO	15,000	$12/mo
ULTRA	75,000	$39/mo
MEGA	300,000	$99/mo

Background removal counts as GPU usage, tracked separately from regular CPU requests.

Full API Reference

Everything is documented at theglitch.app/docs — all parameters, response formats, error codes, and the complete MCP setup guide.

If you're building something with this or have questions about the integration, drop a comment.

DEV Community