Which MCP clients are supported?

Claude Desktop, Cursor, and Hermes connect locally via npx (stdio) with an API key. claude.ai, ChatGPT, and OpenClaw connect to the remote endpoint — sign in with your imageat account when you add the connector. Any MCP-compatible client can use either mode.

Do I need an API key for imageat MCP?

For claude.ai, ChatGPT, and OpenClaw: no — sign in with your imageat account when you connect and click Allow. For Claude Desktop, Cursor, and Hermes: yes — create a key on your Projects page and paste it into your local MCP config.

How do I disconnect imageat MCP?

Remove the connector in your AI client, then revoke the Claude (MCP) or similar key on your Projects page. That immediately stops further usage.

What AI models are available through imageat MCP?

Nano Banana Pro, Nano Banana 2, GPT Image 2, Seedream, and Krea 2 for images; Veo, Kling, Seedance, and more for video — plus 20+ editing tools like background removal, relight, and virtual try-on.

How is imageat MCP usage billed?

Exactly like generating on imageat.com — each call deducts the same credits a generation would normally cost. Check your balance any time with the check-credits tool.

Is there a rate limit on imageat MCP?

Yes, scaled to your plan tier, with standard rate-limit headers on every response. See the full API documentation for the limit table.

imageat MCP for any AI

Generate images, video, and 20+ edits from any MCP client — billed to your imageat credits.

Connect in 1 minute API key (desktop)Docs

Remote connector — sign in with imageat, no API key to copy

Copy the imageat URL

You'll paste this into the custom-connector dialog in the next step.

Remote MCP endpoint

https://mcp.imageat.com/mcp

Settings → Connectors

Add a custom connector, name it “imageat”, paste the URL above, and leave OAuth fields empty.

Sign in & allow

Claude opens imageat in your browser — sign in and click Allow. No API key to copy. Usage is billed to your credits; a “Claude (MCP)” key appears on your Projects page (revoke anytime).

What your AI can do

Generate images

Nano Banana Pro, GPT Image 2, Seedream & more — 1K–4K.

Generate video

Veo, Kling, Seedance — text-to-video and image-to-video.

20+ edit tools

Remove background, relight, virtual try-on, object eraser…

Check credits

Read your balance before running a job.

The entire toolkit, from one prompt

Text-to-image

Describe it, get a finished shot

Ask your agent for a product photo, a portrait, or a scene in plain language. It picks the right model and resolution — no prompt engineering required.

Generate an image

Product photography

Turn a product into a campaign shot

Drop in a reference photo and describe the setting — studio, lifestyle, outdoors. imageat composes a polished, on-brand product image in one call.

Create a product shot

Upscale & enhance

Fix low-res images automatically

Point your agent at a blurry or low-resolution image and it comes back sharp, detailed, and ready for print or web — no manual editing.

Upscale an image

Before

After

Cinematic video

Animate a still into cinematic motion

Start from an image or a prompt and let the agent build camera movement, lighting, and pacing — powered by Veo, Kling, and Seedance.

Generate a video

The latest models, all in one place

Nano Banana Pro

GPT Image 2

Seedream 4.5

Seedance 2.0

Veo 3.1

Kling 3.0

Grok Imagine

Pixverse

Made with imageat models

Fine art style

Architecture

Fantasy

Photography

Headshot

Studio portrait

Product

Multi-step edits

Got any questions left?

We've answered the most frequently asked ones.

It's a Model Context Protocol server that gives AI clients — Claude, ChatGPT, Cursor, OpenClaw, Hermes, and others — tools to generate images, generate video, and run imageat's editing features. Every call is billed against your imageat account credits, same as generating on the site.

Start generating from your AI

claude.ai, ChatGPT, and OpenClaw: add the connector and sign in — no key to paste. Claude Desktop, Cursor, and Hermes: grab an API key from Projects and add it to your config.

See setup steps Get API key (desktop)

Full setup, tools, and error codes in the API documentation.