1
Copy the OpenArt Connector URL
You'll paste this URL into Claude in the next step.
https://mcp.openart.ai/mcp
OpenArt MCP & CLI
Generate Images & Video Inside Your AI Agent
2
Open Connector Settings
In the Claude app or on claude.ai, go to Settings → Connectors and choose Add custom connector. For Team and Enterprise plans, Follow instructions instead.
3
Connect and Allow Access
Click Add → Connect and sign in to OpenArt. Set permissions to Always Allow. Once connected, you can ask Claude to generate images or videos using OpenArt.
1
Install the Plugin
Find the OpenArt plugin here, then click Install Plugin.
2
Authorize
Sign in with your OpenArt account to connect, it only takes a few seconds, and no API keys are needed.
3
Activate & Generate
Type @OpenArt in the message box to enable it, then simply ask ChatGPT to generate images or videos.
1
Open MCP Settings
Settings → MCP & Integrations → New MCP server (opens mcp.json).
2
Add the OpenArt Server
Paste this into mcp.json and save:
{
"mcpServers": {
"openart": {
"url": "https://mcp.openart.ai/mcp"
}
}
} 3
Sign In Once
Click the Needs login prompt next to the server and approve in the browser. OpenArt tools now show up in Agent chats.
1
Install the Plugin
Find 'OpenArt AI' in Kimi plugins store and install.
2
Authorize
Sign in with your OpenArt account to connect, it only takes a few seconds, and no API keys are needed.
3
Activate & Generate
Type /OpenArt in the message box to enable it, then simply ask Kimi to generate images or videos.
1
Find the MCP Settings
Look for Connectors, Integrations, or MCP servers.
2
Point It at OpenArt
Add a server with transport Streamable HTTP and the URL above. No API key, no extra config.
https://mcp.openart.ai/mcp
3
Authenticate via OAuth
Sign in with your OpenArt account on first use. Credits and results stay tied to your account.
1
Install in terminal
macOS and Linux
curl -fsSL https://raw.githubusercontent.com/OpenArt-AI/cli/main/install.sh | sh Windows
irm https://raw.githubusercontent.com/OpenArt-AI/cli/main/install.ps1 | iex 2
Sign in
Login with your OpenArt account.
openart login 3
Generate
Start generate an image or video.
openart generate image "a red fox in the snow" --model nano-banana-2 -o ./out/ 1
Install in terminal
macOS and Linux
curl -fsSL https://raw.githubusercontent.com/OpenArt-AI/cli/main/install.sh | sh Windows
irm https://raw.githubusercontent.com/OpenArt-AI/cli/main/install.ps1 | iex 2
Sign in
Login with your OpenArt account.
openart login 3
Generate
Start generate an image or video.
openart generate image "a red fox in the snow" --model nano-banana-2 -o ./out/ 1
Install in terminal
macOS and Linux
curl -fsSL https://raw.githubusercontent.com/OpenArt-AI/cli/main/install.sh | sh Windows
irm https://raw.githubusercontent.com/OpenArt-AI/cli/main/install.ps1 | iex 2
Sign in
Login with your OpenArt account.
openart login 3
Generate
Start generate an image or video.
openart generate image "a red fox in the snow" --model nano-banana-2 -o ./out/ 1
Install in terminal
macOS and Linux
curl -fsSL https://raw.githubusercontent.com/OpenArt-AI/cli/main/install.sh | sh Windows
irm https://raw.githubusercontent.com/OpenArt-AI/cli/main/install.ps1 | iex 2
Sign in
Login with your OpenArt account.
openart login 3
Generate
Start generate an image or video.
openart generate image "a red fox in the snow" --model nano-banana-2 -o ./out/If you are using Claude Code or Codex, it's better to
GitHubWhat OpenArt MCP Brings to Your AI Agent
Connect OpenArt with Claude, ChatGPT, Cursor, and more to unify image generation, video creation, and all your creative library in one seamless workflow.
Create Both Images and Videos
Turn text into images and videos, all through a single OpenArt connection.
Try it Now
Guide Videos with Start & End Frames
Turn selected start and end frames into a polished image to video while keeping the transition under your direction.
Try it Now
Storyboard and Plan with Smart Shot
Hand Smart Shot your references and a text prompt, and it plans the shots, framing and motion for a multi-cut, finished spot.
Try it Now
One Connection, Every AI Model
Add one URL and your agent instantly accesses OpenArt's full catalog of image and video models. Every new model is synced automatically, so your setup never needs manual updating.
Try it NowImage
-
GPT Image 2 Recommended -
Nano Banana Pro - Nano Banana 2
- Nano Banana 2 Lite
-
Seedream 5 Pro - Seedream 5 Lite
- Seedream 4.5
-
Kling 3 Omni
Video
- Seedance 2.5
- Seedance 2.0
- Seedance 2.0 Mini
- Gemini Omni 1.1 Flash
- Gemini Omni Flash
- Kling 3 Omni
-
Grok Imagine 1.5 -
PixVerse V6 -
Wan 2.7 - Smart Shot
Reuse and Reference Your OpenArt Library
Pull any past OpenArt generation into a new creation as a style reference or a video start frame. Your entire creative history is always within reach for reference.
Try it Now
Create a premium beverage ad for the Citra orange soda can, surrounded by fresh oranges and splashing citrus. Use vibrant studio lighting, crisp product detail, and polished commercial styling.
GPT Image 2Seedance 2.516:924s video
Stay Connected to Your Workspace
Work with the same OpenArt account, credits, projects, and workspaces you already use, for maximum convenience.
Try it NowJust Ask. Here's What OpenArt MCP Creates Through Chat
See what becomes possible when OpenArt is connected to your AI agent. Just chat with your agent to create anything!
Generate Images Without Leaving Chat
Generate a moodboard of a desert-glass perfume bottle — 4 variations, studio light, 1:1.
Text-to-image across OpenArt's image models. Your agent returns the results right in the conversation and saves them to your library.
Create Videos Without Leaving Chat
Animate this product shot into a 5-second cinematic ad.
Make videos from text or images directly through your AI agent. OpenArt handles generations and returns the finished result right in the conversation.
Organize Work Into Projects
List my OpenArt projects, then create three new concepts in my Q3 campaign project.
List your OpenArt projects and pick one before you create, so each generation lands in the right place. No manual file shuffling and switching tabs.
Reuse Your Best Creations
Use my last three renders as references.
Pull assets from your OpenArt library and use previous generations to maintain style, characters, and creative consistency.
Stay Focused on Creating
How many credits do I have left?
Check plans, credits, and account information instantly without opening OpenArt settings. All information flows freely between OpenArt and your AI agent.
The Latest AI Models, All in One Place
Frequently Asked Questions
MCP (Model Context Protocol) is an open standard that lets AI agents use external tools. Once connected, your agent can call OpenArt to generate images and video — and return the results directly in your conversation.
In Claude (web, desktop, or mobile), go to Settings → Connectors, click Add custom connector, name it OpenArt, and paste the server URL. Click Connect, sign in with your OpenArt account, and ask Claude to generate an image.
Yes. In Cursor, open Settings → MCP & Integrations → New MCP server, add OpenArt to mcp.json with the server URL, then click the login prompt and approve in your browser. OpenArt tools then show up in Agent chats.
Generations through MCP spend the same credits as if you're generating on openart.ai. Each generation costs credits based on the model and resolution, and your existing plan works without any extra setup.
Everything your agent creates is returned in the conversation and saved to your OpenArt library, so you can keep iterating on openart.ai whenever you want a full canvas.
Yes. In Claude, OpenArt provides an upload widget so you can upload an image to
OpenArt right inside the conversation and use it as a reference; you can also
paste an image URL. Pasting an image URL works in any client that supports it.
OpenArt right inside the conversation and use it as a reference; you can also
paste an image URL. Pasting an image URL works in any client that supports it.
It lets you browse, reuse, and create assets inside the active OpenArt workspace instead of treating every generation as a disconnected file. For teams, it also helps route work to projects and keep credit usage tied to the right workspace.
Just mention OpenArt in your prompt so your agent knows where to route the request. For example: "Use OpenArt to generate a 5-second product video" or "With OpenArt, create 4 image variations of this concept." This keeps the agent from guessing and makes results more predictable. Once OpenArt becomes your go-to for image and video, you can drop the name entirely.
No, the CLI install command needs to run in a coding agent's desktop app (like the terminal built into Claude Code, Cursor, etc.) or a local terminal on your computer, it can't be run from a web browser.
OpenArt CLI works on macOS, Linux, and Windows. You can use it with Claude, ChatGPT, Cursor, and other AI agents. It’s ideal for terminal-based workflows, automation, and batch creation.
Go to Settings → Apps & Connectors → Advanced, enable Developer Mode, then create a connector named OpenArt with the server URL and authorize via OAuth. In the chat, type @ and select OpenArt to enable it, skipping this causes ChatGPT to generate images on its own instead of using the connector. Then ask ChatGPT to generate an image; to use a reference image, paste an image URL in your prompt.
No. Add the server URL in your agent's settings and sign in with your OpenArt account through OAuth. There are no keys to create, rotate, or leak.
Generate images and video, discover available models, track results, upload reference files, check your account and credits, switch workspaces, and organize projects, all from the conversation. The tool list updates automatically as new tools ship.
Image models include Nano Banana 2, Nano Banana Pro, GPT Image 2, Seedream 4.5, and Seedream 5 Lite. Video models include Kling 3 Omni, Grok Imagine 1.5, Seedance 2.0, Seedance 2.0 Fast, Wan 2.7, and PixVerse V6. Your agent can pick a model automatically or use one you name, and new models appear automatically as OpenArt adds them.
Yes. On managed workspaces (Claude Team/Enterprise or ChatGPT Business/Enterprise), an Owner or admin adds the OpenArt connector for the organization first. In Claude, use Settings → Organization settings → Connectors → Add and enter the server URL https://mcp.openart.ai/mcp. In ChatGPT, that means enabling developer mode and creating the connector. This only needs to be done once, after it's added, OpenArt becomes available to everyone on the plan, and each member simply signs in with their own OpenArt account via OAuth. On personal paid plans (Claude Pro/Max, ChatGPT Plus/Pro) you can add it yourself. Once connected, generations spend the workspace's credits and results save to the workspace library so your team can pick up where you left off.
Image generation, video creation, image-to-video, external image references, and your OpenArt library are all available today. Character and Smart Shot are on the roadmap.
Install the CLI, run openart login, and start creating from your terminal. See the complete setup guide on OpenArt CLI GitHub.
Check our email announcements or our GitHub/MCP page, both are kept current with new feature releases and changelogs.