Sanity MCP. Control your entire CMS from chat, not a UI.
Sanity provides full control over your headless CMS content by letting any AI agent perform operations through natural conversation. This MCP lets you query documents using custom GROQ queries, create new drafts, update existing fields, and even manage entire datasets without ever opening the Sanity Studio interface. You can list all project members or browse media assets—it turns complex backend administration into simple chat commands.
Give Claude and any AI agent real-world access
Ask the agent to search and retrieve documents by specific criteria using advanced GROQ query syntax.
Add new documents or modify existing fields (like changing a status from draft to published) with single commands.
Browse all available document types within your project scope.
List existing datasets or create new ones, isolating content for different stages like staging or development.
Browse your image library and audit who has access to the project's backend tools.
Ask an AI about this
Waiting for input…
What AI agents can do with Sanity Alternative: 12 Tools
These tools let you perform every common task in headless CMS management—from drafting content to auditing user access—all through your AI chat interface.
Make your AI actually useful.
Add this MCP to Claude, Cursor, or Windsurf and your AI stops guessing. It gets real tools to look things up, take action, and handle the stuff you keep doing by hand.
Start using Sanity MCPCreate Dataset
Creates an empty, isolated content collection for testing purposes (e.g., 'staging').
Create Document
Generates and saves a brand new document with specified type fields.
Delete Document
Permanently removes an existing content item from the database; this action cannot...
Get Document
Retrieves all field data for a single, specific document using its unique ID.
List Datasets
Displays all the isolated content environments currently configured in your project.
List Documents
Lists every available document of a specific type within the current dataset, including metadata.
List Image Assets
Retrieves metadata for all images and media files stored in your content library.
List Users
Lists every user account that has access to the project, useful for security audits.
Query Documents
Runs highly customized GROQ queries across your content to find data based on...
Search Documents
Performs a broader search of documents, optionally filtering by type and setting a...
Update Document
Modifies specific fields (like title or status) on an existing document without...
Security and governance baked right in.
Pick your AI client below to get set up. Just create a Vinkius account, subscribe, and you're instantly up and running. We handle the entire backend infrastructure, delivering out-of-the-box support for HTTPS Streamable, SSE, and OAuth2—zero messy routing required.
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on each call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with Sanity, then connect any of our 5,200+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 5,200+ others, all in one place
- Add new capabilities to your AI anytime you want
- Connections are secured and governed automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog weekly
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Sanity. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS CLOUD
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on each call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
The Pain of Context Switching
Today, managing content requires constant context switching. To update a single product page, you might have to jump from your writing app into the CMS UI, find the correct document ID in one tab, run a query in another, and finally navigate back out to save the changes. This clicking dance is slow and error-prone.
With this MCP, that entire process collapses. You just speak to your agent: 'Update product ABC's price to $299.' The agent handles finding the document ID, making sure you're in the right dataset, updating only the price field, and confirming it—all without you ever leaving your chat window.
Content Management with Sanity
The ability to run a complex search is massive. Instead of manually building a filter or running multiple searches for specific document types, the agent executes a precise `query_documents` command based on natural language input.
This means you're not just querying data; you’re getting actionable intelligence instantly. Your AI client has full control over your content lifecycle—it moves beyond simple retrieval to become a true operational partner.
What Sanity MCP does for your AI
Your AI agent now acts as a dedicated content operations assistant for your Sanity project. Instead of navigating multiple dashboards to find a document ID or remembering precise query syntax, you just tell your client what you need done.
It handles everything from drafting new blog posts and updating product details to running complex data checks across different datasets. Need to know which users have access? Or maybe you're staging content and need to create a clean dataset for development testing? Your agent manages it all, keeping track of your entire content lifecycle in one place.
Connecting this MCP through the Vinkius catalog gives any compatible AI client instant, conversational access to your entire content graph. You manage data with plain language commands; the system handles the complex API calls.
019d847b-2e55-70da-9820-8e8800be4293 How to set up Sanity MCP
The bottom line is that your AI client treats complex content management like a simple conversation with an expert assistant.
Subscribe to this MCP, then provide your Sanity Project ID, Dataset name, and API Token.
Connect your preferred AI client (like Cursor or Claude) to the Vinkius catalog. This gives your agent access credentials.
Tell your agent what you need—for example, 'List all drafts of product documents'—and it executes the necessary commands.
Who uses Sanity MCP
This MCP is for technical content teams, developers building data-driven sites, and senior editors who spend too much time jumping between UI screens. It’s perfect for anyone who gets frustrated when a simple task requires navigating five different tabs just to run one query.
Searches for specific articles by keyword, reviews the full details of existing documents, and manages media assets without leaving their writing interface.
Runs complex GROQ queries to validate data structures or programmatically creates new document drafts with required schemas for testing purposes.
Audits project access by listing all users and manages the creation of isolated datasets, ensuring development content doesn't mess up production data.
Benefits of connecting Sanity MCP
You eliminate the need to manually run complex queries. Instead of writing intricate GROQ code every time you need data, simply ask your agent to query documents using query_documents and get instant results.
Content publishing becomes conversational. You can draft content or change a document's status from 'draft' to 'published' by asking the agent to run an update_document command, skipping clicks entirely.
Data isolation is effortless. If you need to test changes without affecting live data, use the MCP to list and create dedicated datasets via list_datasets and create_dataset.
Auditing access used to mean clicking through user roles. Now, you simply ask your agent to run list_users, getting a clean roster of who has access instantly.
Media management is direct. Instead of navigating the asset library manually, you can list images and their metadata by running list_image_assets directly from your chat window.
Sanity MCP use cases
Finding all articles needing a content review
A Content Editor needs to find every product post that was written last week but hasn't been assigned an image yet. They ask their agent to run a query, and the MCP executes query_documents using filters for 'product' type and date criteria, returning only the IDs of articles needing review.
Setting up a clean environment for QA
A Developer is working on a new feature and needs to test content against production data. They first ask their agent to list_datasets to verify 'production' exists, then command it to create_dataset named 'qa-test', ensuring zero risk to live content.
Mass updating product details
A Marketing Manager needs to change the status of 15 old blog posts from draft to published. Rather than opening 15 documents, they instruct their agent to update_document for a list of IDs and set the 'status' field to 'published'.
Retrieving specific document details by ID
An Editor gets an email with only a content item ID. Instead of guessing where it lives, they ask their agent to get_document using the provided ID and instantly retrieve all associated fields for review.
Sanity MCP tradeoffs
What to watch out for, and the recommended way to handle each one.
Trying to manually list documents
The user tries to guess if a document exists by running multiple, slightly different search queries across the API.
Instead of guessing, tell your agent to list_documents with the specific content type you are looking for (e.g., 'post'). This gives you a structured list of all existing records.
Overwriting critical fields
A developer runs an update command without specifying which field needs changing, potentially wiping out valuable metadata.
Always use update_document and provide a JSON object that specifies only the exact fields you want to modify. The agent ensures no other data is accidentally changed.
Confusing datasets
A team member makes changes in the wrong environment, inadvertently publishing test content directly to the live site.
Before making any write operation, always use list_datasets and confirm you are operating against the intended dataset (e.g., 'development') by using create_dataset first.
When to use Sanity MCP
Use this MCP if your primary bottleneck is content workflow or data manipulation. Specifically, if you need to perform CRUD operations on headless CMS content—creating drafts, modifying fields, deleting old articles—then this is the tool. It lets your agent act as a full-stack content editor.
Don't use it if all you need is simple data retrieval for display purposes (e.g., 'give me the title and slug of every post'). In that case, a read-only API connector might be simpler and faster to integrate. But because this MCP handles writes (create_document, update_document) alongside reads (query_documents), it's ideal when your task requires making changes based on data discovery.
If your goal is just asset management without content, you probably need a dedicated media library tool instead of one that includes document creation.
Frequently asked questions about Sanity MCP
How do I run custom searches with Sanity using the Sanity MCP? +
You use the query_documents tool. You simply ask your agent what data you need, and it translates that into a powerful GROQ query to pull exactly what you're looking for.
Can I create content drafts with Sanity using this MCP? +
Yes, you use the create_document tool. You can specify the document type and fill in basic content fields (like a title) to generate a new draft ready for editing.
Is it safe to delete documents using the Sanity MCP? +
The delete_document action is irreversible. Always confirm with your agent that you have the correct document ID and are sure you want to permanently remove the content before confirming.
What if I need to test changes in a sandbox environment? +
You should use list_datasets first, then run create_dataset to make an isolated copy. This ensures any testing or mistakes won't affect your live production content.
How do I check who has access to my Sanity project? +
Simply ask the agent to execute the list_users tool. It will retrieve a list of all registered team members and their current permissions within the project.