Give your AI agents eyes & imagination.

One MCP server that lets any AI agent generate, edit, inpaint, remove backgrounds, upscale, and transform images using the world's best image models.

$npx skills add imagemcp/skill
mcp-server
STDIO/SSE READY
POST /mcpgenerate_image
{
  "tool": "generate_image",
  "prompt": "Cyberpunk cityscape at dusk...",
  "model": "flux-pro"
}
200 OKimage url returned
MCP Spec v1.0
Connects with Claude, Cursor, ChatGPT & custom agents
Tool-Call Ready

Available for these AI Agents & IDEs

The Problem

The visual creation gap in AI stacks

AI agents handle text easily, but visual creation, inpainting, background removal, and upscaling across providers require fragmented glue code.

Fragmented API Integrations

Different APIs, auth schemes, and payload formats for every provider.

Editing & Inpainting Friction

Complex mask pipelines, background cutouts, and base64 handling.

Unwanted Operations Overhead

Managing rate limits, retries, and API key sprawl across services.

AgentEditorBotModel AModel BModel C
The Toolbelt

Eight tools. One connection.

This is the entire surface your agent gets after a single config line — no per-provider SDKs, no separate keys.

8–30 cr

generate_image

Text-to-image, or image-to-image when you pass a reference.

promptmodel?aspectRatio?referenceImages?
16–38 cr

generate_transparent_image

Generate and cut out in one call — returns a transparent PNG.

promptmodel?aspectRatio?
10 cr

edit_image

Inpaint, retouch or restyle an existing image from a prompt.

imagepromptmodel?
8 cr

remove_background

Isolate the subject and return a clean alpha cutout.

image
15 cr

upscale_image

Super-resolution up to 4K with detail preserved.

imagescale?
5 cr

text_to_svg

Real vector output — logos, icons and shapes as editable SVG.

promptstyle?
1 cr

compress_image

Shrink file size with quality control and format targeting.

imagequality?targetFormat?
1 cr

convert_format

Move between PNG, JPG, WEBP, GIF, BMP and SVG.

imagetargetFormat

Credit cost varies by the model your call routes to. Every call lands in your request log with its exact cost.

Developer First

One config. Generate & Edit any image.

Add the server to any MCP client, or call the API directly. Synthesize new images or modify existing visuals, inpaint masks, remove backgrounds, and upscale through unified tool calls.

  • Unified schema for generate_image & edit_image
  • Supports Inpainting, Outpainting, Background Removal & 4K Upscaling
  • Automatic routing, retries & fallbacks across 40+ models
  • Works with Claude, Cursor, ChatGPT, Antigravity & terminal CLI
Tool:
generate.ts
import { Client } from "@modelcontextprotocol/sdk";

const mcp = new Client({ server: "imagemcpserver.com" });
await mcp.connect();

// Supports text-to-image AND optional input image guidance
const { image } = await mcp.call("generate_image", {
  prompt: "A serene Japanese garden at dawn, soft mist",
  image:  "https://example.com/reference.jpg", // optional input image
  model:  "google/gemini-2.5-flash-image",
  aspectRatio: "16:9",
});
How It Works

Four steps. Five minutes.

Connect your AI agent or editor in minutes to generate, edit, and upscale visuals seamlessly.

⚡ Simple 1-click configuration
1

Connect

Add one server URL to your MCP client config.

2

Request

Ask to generate or edit an image via prompt or JSON.

3

Generate & Edit

AI synthesizes, inpaints, retouches, or upscales pixels.

4

Receive

Get high-resolution visual outputs in seconds.

Centralized Hub

Manage everything from your Dashboard

Issue and revoke keys per agent, watch credit spend as it happens, audit every model call, and cap your monthly budget.

imagemcpserver Console

Manage API Keys

Key NameSecret KeyCreatedActions
Production Agentsk-img-pro-••••7a8b2 days ago
Staging Clientsk-img-dev-••••9f2c1 week ago
Use Cases

Built for every AI workflow

If your software can call MCP, it can now see.

Autonomous Agents

Research, web-search, and design agents can synthesize new artwork or edit existing images mid-task — no manual photo editor needed.

agent.call("edit_image", { image: photo, prompt: 'inpaint cyberpunk skyline' })

AI Editors & IDEs

Add image generation, inpainting, and 4K upscaling to any editor with standard MCP config.

Content Automation

Blog visuals, social graphics, background removal, and product retouching — on autopilot.

Game & UI Assets

Generate and edit sprites, textures, asset variations, and concept art on demand.

Website Generation

Create hero images, vector SVGs, transparent graphics, and landing page visual assets directly during website build workflows.

Internal Tools

Reliable image generation for your team's dashboards and ops tools — with one key, one bill, and full audit logs.

Per-key audit logsMonthly spend capRevoke without redeploying
8MCP tools built in
40+Image models routed
20+Agents & IDEs supported
30Free credits / month
Why we built it
Every agent framework already knows how to think and how to write. Almost none of them can put a pixel on a canvas without a week of vendor glue code. That's the gap this server closes — in one config line.

One key, every model

Swap between Gemini, FLUX, Recraft, Imagen or SDXL without touching your agent code or juggling provider accounts.

Every call is auditable

Prompt, model, latency and credit cost land in your request log the moment an agent fires a tool.

Spend can't run away

Credits are pre-paid and per-key. An agent stuck in a loop burns your balance, never a surprise invoice.

Pricing

Start free. Pay when it ships.

Prepaid credits, one bill for every model. No seat pricing, no minimum commitment, cancel from the dashboard whenever you like.

Free Tier

Free

$0/mo

30 credits / month

  • 30 Credits / month
  • Basic image editing tools
  • Text → Image
  • API & MCP Server access
Start free

Starter

$20/mo

1,000 credits / month

  • 1,000 Credits / month
  • All image editing tools
  • Text → Image
  • Text → SVG
  • Image → SVG
  • API & MCP Server access
Choose Starter
Most Popular

Pro

$50/mo

3,000 credits / month

  • 3,000 Credits / month
  • Everything in Starter
  • Priority processing
  • Early access to new tools
Choose Pro
Enterprise Scale

Enterprise

$100/mo

8,000 credits / month

  • 8,000 Credits / month
  • Everything in Pro
  • Maximum rate limits & Priority support
Choose Enterprise

Need more headroom or your own provider keys? Compare every plan and credit pack.

FAQ

Questions before you wire it in

Still stuck? Mail us at hello@imagemcpserver.com and a human will answer.

MCP endpoint
https://mcp.imagemcpserver.com/mcp

Model Context Protocol is the open standard AI clients use to discover and call external tools. Point a client at this server once and your agent gains eight image tools it can call on its own — no plugin code, no SDK per provider.

Give your agents pixels today

One server, every major model. Wire it into Claude, Cursor, ChatGPT or your own stack in about five minutes.

No credit card required. Free tier includes 30 credits every month.