Cosmic MCP for AI. Control your entire CMS via natural conversation.
Works with every AI agent you already use
…and any MCP-compatible client








Connect to your AI in seconds.
Cosmic (Headless CMS & Content API) MCP. Control your content infrastructure, media library, and data structures directly through natural language.
This tool lets you build site content, manage image assets, and update object types without ever touching a database console.
What your AI can do
Batch objects
Applies a set of changes to multiple objects (up to 25) at once.
Create media
Uploads and creates new media assets, like images or videos, in the bucket.
Create object
Builds a brand-new content object with specified metadata.
Create completely new pieces of structured content (like a blog post or product listing) with specific metadata.
Fetch specific content items, list multiple results, or check an object's full revision history by ID.
Upload new images and videos, delete old ones, or update the metadata attached to existing files in the library.
Programmatically create, read, and update the blueprints that define what kind of data you can store (the object types).
Apply changes to up to 25 different objects in a single request for high-efficiency maintenance.
Ask an AI about this
Waiting for input…
Cosmic (Headless CMS & Content API) Tools (21)
These tools give your agent direct access to every core function of a modern headless CMS, from creating content drafts to managing media assets.
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 Cosmic (Headless CMS & Content API) on VinkiusBatch Objects
Applies a set of changes to multiple objects (up to 25) at once.
Create Media
Uploads and creates new media assets, like images or videos, in the bucket.
Create Object
Builds a brand-new content object with specified metadata.
Create Object Type
Defines and creates an entirely new structure for content objects in the CMS.
Create Revision
Adds a specific historical checkpoint or version to any existing object.
Delete Media
Permanently removes media files from the content library.
Delete Object
Removes a specific, existing content object.
Delete Object Type
Deletes an entire template or blueprint for objects in the CMS.
Get Agent Conversation History
Retrieves past messages and actions from a specific AI agent conversation history.
Get Object
Fetches the full data record for one object using its unique ID.
Get Object Type
Retrieves the specific definition or blueprint for a content type by its slug.
Get Revision
Gets the data snapshot of an object at a particular historical point.
List Agent Conversations
Lists all past conversations associated with a specific AI agent user account.
List Media
Shows a list of every media asset currently stored in the bucket.
List Object Types
Retrieves a catalog of all existing content type templates used across the CMS.
List Objects
Lists multiple objects and allows you to filter results using structured query...
List Revisions
Retrieves a list of all available historical versions for a single object.
Send Agent Message
Sends a new message directly to an AI agent account, triggering a response.
Update Media
Changes metadata—like alt text or descriptions—for existing media files.
Update Object
Modifies the content and metadata of an object that already exists.
Update Object Type
Makes changes to the blueprint or structure of a content type template.
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 every call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with Cosmic (Headless CMS & Content API), then connect any of our 5,100+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 5,100+ others, all in one place
- Add new capabilities to your AI anytime you want
- Every connection is secured and compliant automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog every week
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Cosmic. 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 INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on every call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
Works with Claude, ChatGPT, Cursor, and more
The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.
This connection provides 21 powerful capabilities that interface natively with Claude, ChatGPT, Cursor, and other compatible AI platforms. No middleware. No custom integration required.
The old way of updating site content feels like a chore.
Today, changing anything about your website means jumping through hoops. You copy the text from one dashboard into another, click 'save drafts' in the CMS, then manually go to an asset manager tab to find and swap out images. It’s a messy process of clicks and tabs that breaks focus.
With this MCP, all that friction vanishes. Your agent handles the entire content lifecycle—from drafting the object using `create_object` to swapping out media assets with `update_media`. You just talk to it; it does the rest.
Controlling Object Types and Schema
Before, if you needed a new kind of content—say, 'Staff Profile' instead of 'Blog Post'—you had to wait for an engineer to manually create the template structure in the admin panel. It was slow and required specific permissions.
Now, your agent can execute `create_object_type` on demand. You define the new blueprint right through conversation, allowing product teams to iterate on their content models immediately.
What your AI can actually do with this
This connector gives you full control over any modern headless CMS setup. You can treat your entire website's content—the articles, the product listings, the media files—as code that your AI agent controls. Instead of logging into a separate web interface and clicking through menus to make changes, your agent handles everything in conversation.
Need to draft a new blog post and update its metadata? Done. Want to delete old images or check which content types exist? No problem. The ability to manage object structures, media assets, and revisions means you're treating your CMS like another service layer right alongside your backend code. Because Vinkius hosts this MCP in the catalog, you can connect it once from any AI client and access all these content operations instantly.
019e5d0d-48fc-7365-98d1-00ee19b863e8 Here's how it actually works
The bottom line is that you give your AI client the necessary keys and context; it handles all the complex CMS calls for you.
First, subscribe to this MCP and provide your Cosmic Bucket Slug, Read Key, and Write Key credentials.
Next, point any AI client—like Cursor or Claude—to the Vinkius catalog connection. Your agent now has access to all content tools.
Finally, ask your agent to perform a task: 'Create an object called X with this metadata,' or 'List all media assets in the bucket.'
Who is this actually for?
This is for developers who write content-driven apps, product managers who need to automate publishing workflows, or technical content writers tired of jumping between a CMS dashboard and their IDE.
Needs to query the system structure to ensure new components correctly read data types from the CMS before writing code.
Uses the AI agent to draft a series of articles, upload associated images, and organize all the related metadata without leaving their workspace.
Runs scheduled tasks via automation agents to check for outdated content types or perform batch updates across dozens of object records.
What Changes When You Connect
You don't have to switch context. With this MCP, you can draft a new object and immediately check its associated media library using list_media—all from your agent chat window. It keeps the workflow contained.
Avoid data silos by automating updates. Need to update 20 product descriptions? Use batch_objects to push changes across multiple records in one go, saving hours of manual API calls.
Maintain content integrity with full visibility. Check object history using get_revision or review the past actions via list_revisions. You always know who changed what and when.
Build flexible data models on demand. If you need a new kind of data (say, 'Testimonials'), use create_object_type to build its blueprint before content even gets written.
Manage media alongside text. Use update_media to change an image's alt tag immediately after creating the accompanying object using create_object.
See it in action
A marketing team needs to roll out a campaign landing page.
The manager asks their agent: 'Draft five new articles about Q3 earnings, using the 'news' object type.' The agent responds by executing create_object for all five drafts and then calls list_media to confirm that the required hero banner is present.
A developer needs to update metadata across thousands of records.
Instead of writing a complex script, they tell their agent: 'Update the SEO title for all objects in the 'product' type that haven't been touched this month.' The agent uses update_object and list_objects to execute the mass change.
A product team needs to audit old content.
The PM asks: 'Show me all objects of type 'legacy-product' that have never been viewed in 90 days.' The agent runs list_objects with a complex query and flags the records for archival.
A tech writer needs to ensure new content adheres to standards.
The writer tells the agent: 'I need to create an object, but first check if the required image type exists.' The agent uses list_object_types and then calls create_object only after validation.
The honest tradeoffs
Trying to update everything with one call
The user tries to tell the agent: 'Update this object, change its image, and also create a new type.' This fails because update_object doesn't handle schema changes.
You must separate these actions. First, use create_object_type to define the new blueprint. Then, use update_object for content changes, or create_media for image uploads.
Forgetting to check object existence
The user asks the agent to delete an object by name, but it fails because names aren't unique identifiers. The API needs a specific ID.
Always start by using get_object or list_objects to verify and retrieve the correct Object ID first. Never assume the ID.
Overwriting media instead of updating it
The user uploads a new image and tells the agent to 'update the banner.' This might delete the old metadata or fail if the asset wasn't properly linked.
Use update_media when you only want to change the alt text or description, or use create_media if it is a brand new file.
When It Fits, When It Doesn't
Use this MCP if your content workflow relies on managing structured data—meaning you need to create, update, and track objects that have specific metadata fields. You need to control the blueprint (object types) and the assets (media). Don't use it if you just need a simple database query or direct row access; for those scenarios, you need a raw database connector type tool instead. Use this MCP when your primary goal is content governance and structured publishing. If the task is purely about sending messages between agents, stick to send_agent_message; don't try to use object tools for that.
Questions you might have
How do I list all available media with the Cosmic API? +
You use the list_media tool. This function shows every asset currently in your bucket, letting you find specific files or count how many assets are ready for deployment.
What if I want to change an object's metadata but not its content? +
Run update_object. This tool lets you modify the surrounding data—like changing a 'featured' flag or adjusting the last modified date—without touching the core body text.
Can I view how an object looked six months ago using get_revision? +
Yes. The get_revision tool lets you pull a historical snapshot of any content object, letting you audit past versions before making changes today.
What is the best way to update multiple objects at once? Should I use update_object repeatedly? +
No. Use batch_objects. This tool lets you group up to 25 object updates into a single call, which is much faster and more efficient than running 25 separate operations.
When using `list_object_types`, how do I see all the available content structures in my Cosmic bucket? +
It lists every defined object type in your connected bucket. This allows you to query which specific content schemas exist, helping you target your data operations accurately before attempting to read or write objects.
I need to remove an old schema; what is the process for using `delete_object_type`? +
You execute delete_object_type with the specific slug of the type you want gone. Be careful, because this action permanently removes the structure and prevents future objects from being created under that name.
When using `list_objects`, what format do I provide for the query parameter to filter content results? +
You must pass a URL-encoded JSON object within the query parameter. This lets you precisely limit the list of objects based on specific metadata fields, rather than retrieving everything in the bucket.
If I only want to change an image's title or tags, should I use `update_object` or `update_media`? +
Use update_media. This tool lets you manipulate the metadata associated with a media asset—like changing its description or adding tags—without needing to re-upload the physical file itself.
Can I filter objects by a specific type like 'blog-posts'? +
Yes! Use the list_objects tool and provide a JSON query string like {"type":"blog-posts"} in the query parameter to filter your results.
How do I update the content of an existing object? +
You can use the update_object tool. Provide the Object ID and the new values for title, metadata, or status to modify the entry.
Is it possible to see who made changes to an object? +
Yes, you can use the list_revisions tool for a specific object to see its history, and get_revision to inspect the details of a specific version.
We've already built the connector for Cosmic. Just plug in your AI agents and start using Vinkius.
No hosting. No infrastructure. No complex setup.
All 21 tools are live and waiting.
You're up and running in seconds.
Vinkius gives your AI agents access to the full catalog of app connectors, all fully managed, secure, and enterprise-ready. One subscription, every tool you need.
Built, hosted, and secured by Vinkius. You just connect and go.