What MCP enables
With MCP connected, AI assistants can:- Discover your design system — search components, understand their properties and purpose
- Find content types — see what structures are available and their schemas
- Create content — generate new content that conforms to your content types
- Update content — modify existing drafts with new text, images, or components
- Search assets — find images and media to use in content
- Maintain compliance — all content respects your defined structures and allowed components
Supported AI tools
Metabind works with:| Tool | Setup |
|---|---|
| Claude | Setup guide (browser, desktop, terminal) |
| Cursor | Setup guide or install now |
| GitHub Copilot | Setup guide |
Connect your AI tool
1
Open your AI tool's MCP settings
Each tool has its own way to configure MCP servers. Consult your tool’s documentation for where to add server configurations.
2
Add the Metabind server
Add Metabind as an MCP server:
3
Authenticate
When you first use a Metabind tool, you’ll be redirected to authorize access. Sign in to Metabind and approve the requested permissions.
4
Start creating
Once connected, you can start working with natural prompts. For example:
“Create a new article in my Acme Corp project about our Q4 product updates. Use the press release hero image.”The AI will find your project, discover the Article content type, locate the image, and create a draft. Changes sync live to Metabind, so you can watch the content take shape in real time.
Example workflows
Create an article
Ask your AI assistant something like:“Create a new article about our product launch using the Article content type. Use the hero image from assets tagged ‘product-launch’.”The AI will:
- Search for the Article content type and get its schema
- Find the hero image in your assets
- Create a new draft article with the content filled in
Update existing content
“Update the ‘Welcome’ content to mention our new features. Keep the existing structure but refresh the body text.”The AI will:
- Find the existing Welcome content
- Understand its current structure
- Update the text while preserving the component structure
Explore the design system
“What components are available for building marketing pages?”The AI will search your components and describe what’s available, helping you understand your design system options.
Semantic search
Metabind understands natural language, so you can describe what you’re looking for instead of using exact keywords:- “What content types are available for showcasing products?” — discovers product-related types
- “Find images from the office” — finds team photos
- “What components do we have for calls to action?” — locates CTA buttons
Governance
AI assistants work within your established guardrails:- Draft by default — AI-created content starts as drafts for human review
- Schema validation — content must conform to your content type structures
- Type compliance — only valid content types and allowed components can be used
- Scoped permissions — you control what each AI tool can access
- Audit trail — track who created or modified content