Visiojoy MCP
Generate AI images & videos in Claude, ChatGPT and Cursor
Connect once and ask your assistant for product shots, edits and short videos. 12 image models and 7 video models, billed to your Visiojoy credits. No API key.
Claude will show a safety warning and ask you to confirm. Check that the name is Visiojoy and the URL is https://mcp.visiojoy.com, then click Continue.
Connect your assistant
Add Visiojoy to your assistant, then sign in with your Visiojoy account. New accounts start with free credits.
Server address
https://mcp.visiojoy.comClaude
Web, desktop and mobile
- Click Connect to Claude below. The Add custom connector form opens with Visiojoy filled in.
- Click Continue, keep the detected sign-in options, and sign in to Visiojoy.
Claude will show a safety warning and ask you to confirm. Check that the name is Visiojoy and the URL is https://mcp.visiojoy.com, then click Continue.
Or add it yourself: Settings → Connectors → Add custom connector, then paste the server address above.
Official: Anthropic's custom connector guide
ChatGPT
Web, Business, Enterprise and Edu plans
- Open ChatGPT Plugins, click +, then Add custom MCP server.
- Name it Visiojoy, enter https://mcp.visiojoy.com/mcp and choose OAuth.
- Click Create as a plugin, install it and sign in to Visiojoy.
- In a chat, type @ and pick Visiojoy.
OpenAI decides which plans can add custom MCP servers. Currently Pro accounts get read-only access, so generating images and videos needs a Business, Enterprise or Edu workspace. Plans and menus change, so check OpenAI's guides for the latest.
Official: OpenAI's setup guide · OpenAI's latest plan rules
Claude Code
Terminal
- Run the command below.
- Type /mcp, pick visiojoy and sign in.
claude mcp add --transport http visiojoy https://mcp.visiojoy.comCursor
And other JSON-configured clients
- Add the server to your MCP settings file.
- Sign in when the browser window opens.
{
"mcpServers": {
"visiojoy": { "url": "https://mcp.visiojoy.com/mcp" }
}
}Works with any other MCP client that supports remote servers with OAuth sign-in. Technical docs
Just ask in plain words
Open a prompt in Claude or copy it into any connected assistant. It picks the right tool, tells you the cost and returns the result in the chat.
Product shots
“Create a clean studio product photo of a ceramic coffee mug on a marble counter with soft morning light, 4:5 for Instagram.”
uses generate_imageTry in Claude
Edit your own photo
“Here's my product photo: [paste image link]. Put it on a sunny beach backdrop and keep the label unchanged.”
uses generate_imageTry in Claude
Bring a photo to life
“Turn this photo into a 5-second video of the product slowly rotating on a turntable: [paste image link]”
uses generate_videoTry in Claude
Short social video
“Make a 5-second vertical 9:16 clip of waves rolling onto a beach at sunset for a TikTok intro.”
uses generate_videoTry in Claude
Clean up an image
“Remove the background from this image, then upscale the cutout: [paste image link]”
uses remove_backgroundTry in Claude
Plan before you spend
“Which video model is cheapest for a 10-second clip, and how many Visiojoy credits do I have left?”
uses list_modelsTry in Claude
Why use Visiojoy through MCP
No API key
Sign in with your Visiojoy account in the browser. Nothing to create, copy or rotate.
Every top model, one connection
GPT Image, Nano Banana, Seedream, Flux, Seedance, Wan, MiniMax and Grok Imagine behind a single server.
Spending guardrails for video
Videos default to the cheapest settings, and any call above 300 credits is stopped until you agree to the cost.
Check before you spend
Your assistant can look up prices and your balance for free before it generates anything.
One account, one library
Same credits as the website, and everything you make shows up in your Visiojoy Assets.
Pay only for results
Credits are held when a task starts and returned automatically if it fails.
The models behind it
Open a model to see examples and what it is best at.
Nano Banana 2.1Quality at speedNEW
GPT Image 2.5Flare drafts, Sunburst polishNEW
Seedream 5 ProStructural fidelity and textNEW- Nano Banana 2 LiteLow-cost iterationNEW
Nano Banana ProInstruction fidelity- Nano Banana 2Fast iteration
- GPT Image 1.5Text rendering and edits
- GPT Image 2Draft-to-asset workflow
- Seedream 5 LiteCost-efficient generation
- Seedream 4.5Cinematic realism
Flux 2 ProColor and layout precision
What your assistant can do
Seven tools: four that create, three free ones that look things up.
generate_imageCreate an image from text, or edit images you provide.
generate_videoCreate a video from text, or animate a photo.
remove_backgroundCut out the subject with a transparent background.
upscale_imageMake an image sharper and higher resolution.
list_modelsLook up models and prices. Free.
get_credit_balanceCheck your remaining credits. Free.
get_task_resultFetch a video, or an image that is still processing. Free.
MCP questions
It's a remote MCP server that lets AI assistants such as Claude, ChatGPT and Cursor generate and edit images and videos with your Visiojoy account. You ask in plain words; the assistant calls Visiojoy and returns the result in the chat.
Claude (web, desktop and mobile), ChatGPT (web, on Business, Enterprise and Edu plans), Claude Code and Cursor, plus any other MCP client that supports remote servers with OAuth sign-in. The server address is https://mcp.visiojoy.com.
No. When you connect, a browser window opens and you sign in with your Visiojoy account. There's no key to create or store.
It uses your Visiojoy credits, at the same prices as the website. Images start at 3 credits, background removal and upscaling cost 1 credit per image, and videos are priced per second. New accounts get 12 free credits. See pricing for plans and credit packs.
Not without your say-so. Videos default to each model's cheapest resolution and default length, which costs 18–140 credits. Any call above 300 credits is refused with the exact cost, so the assistant has to check with you first. Failed tasks are refunded automatically.
12 image models, including GPT Image, Nano Banana, Seedream and Flux, and 7 video models, including Seedance, Wan, MiniMax and Grok Imagine. Some image models need a purchase or subscription. Browse image models and video models.
Yes. Give your assistant a public image link, or an image it can pass along directly, to edit it, animate it or remove its background. Reference videos and audio aren't supported through MCP yet.
The assistant shows you download links, and everything is saved to your Visiojoy Assets next to what you make on the website. Results are private unless you ask for them to be public.
Yes, after your first purchase (any subscription or credit pack): images and videos made after it can be used commercially, and images are watermark-free. Before that, results are for personal, non-commercial use and images may carry a small watermark. See the Terms of Service.
Each call makes one image or one video, there's no audio generation, and reference videos or audio aren't supported through MCP yet. The MCP docs list every limit.
Connect once, create from any chat
Add Visiojoy to Claude in one click, or follow the steps for ChatGPT, Claude Code and Cursor.