Connect AI to your content.
Give AI assistants direct access to your content with Contismo's hosted MCP server. Using the open Model Context Protocol (MCP) standard, AI clients can securely discover, read, create, and update structured content in your CMS.
- Free to start
- No sales call required
- Deploy in minutes
What is MCP?
The Model Context Protocol (MCP) is an open standard that allows AI applications to connect with external tools and data. Rather than manually copying content into an AI conversation, MCP lets assistants access and work with information directly.
Every Contismo project includes a hosted MCP server, giving AI assistants like ChatGPT, Claude, and Cursor access to your content through a standardised connection.
Once connected, AI assistants can explore your content models, find and update entries, work across languages, manage taxonomies, and access UI translations, all within the permissions you define.
How it works
No plugins to install or servers to manage. Enable MCP in Contismo Studio, choose what AI can access, and connect your favourite AI assistant in minutes.
- 1Enable MCP
Open Settings → MCP in Contismo Studio, enable the server, and choose which environment your AI assistant should use.
- 2Choose what AI can access
Select which content models are available and control whether AI can access drafts, unpublished content, and editing tools.
- 3Connect your AI assistant
Connect ChatGPT, Claude, Cursor, or another MCP-compatible client using https://mcp.contismo.com. Authorise securely with OAuth 2.1 or an MCP API key.
- 4Put AI to work
Ask AI to find, create, translate, or update content directly in Contismo, using your existing content models and permissions.
What can you do with MCP?
Once a client is connected, everyday content work becomes a conversation.
Connect your AI assistant
Connect ChatGPT, Claude, Cursor, or any MCP-compatible client to Contismo using a single hosted endpoint. No installation or server setup required.
https://mcp.contismo.comThe easiest way to connect. Add the server URL to your AI client, sign in to Contismo Studio, choose your workspace, and approve the requested permissions. No API key required.
For clients that don't support OAuth, generate an MCP API key in Settings → API Keys. Choose read-only or read/write access, then use the key to authenticate your connection.
Add Contismo to your Cursor MCP configuration using the example below. Replace the placeholder with your MCP API key and adjust the environment if needed.
{
"mcpServers": {
"contismo": {
"url": "https://mcp.contismo.com",
"headers": {
"Authorization": "Bearer YOUR_MCP_API_KEY",
"X-Environment": "master"
}
}
}
}Frequently asked questions
Contismo works with MCP-compatible AI assistants, including ChatGPT, Claude, and Cursor. Simply add your Contismo MCP server URL and follow the connection steps. You can use OAuth for a quick, secure connection or an API key if your client doesn't support it.
Yes! With your permission, AI can create new entries, update existing content, manage translations, and schedule publishing. You decide whether your AI assistant has read-only access or can make changes to your content.
Absolutely. You can choose which content models are available to AI, whether drafts and unpublished content are visible, and whether AI can make changes. These settings are managed directly in Contismo Studio, so you're always in control.
Yes. Your AI assistant can read content in one language, create localized versions for other languages, and work with your project's existing locales. It can also help review translations and identify missing localized content.
No. AI assistants connected through MCP can work with your content, but they cannot change your project's configuration, permissions, team members, or billing settings. These remain under your control in Contismo Studio.
You can revoke access at any time from Settings → API Keys in Contismo Studio. Simply disable or delete the relevant MCP API key, including one created through OAuth. You can also turn off the MCP server entirely to block all connections.
They serve different purposes. GraphQL is designed for delivering content to your websites and applications, while MCP lets AI assistants interact with your content through natural language.
You can use both together. For example, an AI assistant can help create and update articles through MCP, while your website continues to fetch and display them through GraphQL.
