Cosmic MCP. Manage content objects and media assets via conversation.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Cosmic MCP Server gives your AI agent full control over your headless CMS content. You can list, create, and patch objects, audit schema types, and manage media assets directly from chat.
This lets your agent perform complex content operations—like updating a product description and clearing old media—without you ever opening the Cosmic web UI.
It's total content lifecycle management, driven by natural language.
What your AI agents can do
Create cms object
Creates a new CMS object draft by accepting a structured JSON payload.
Get media asset
Retrieves the specific structure and metadata for a given media ID.
Get schema details
Extracts the full list of properties and fields required by an active object type.
Build new content drafts in Cosmic by providing a structured JSON payload.
Get the structural details for a specific media file using its ID.
Examine the required fields and metadata structure for any active object type.
Fetch all data for one specific CMS object ID, logging the full payload.
List all content types and routing spaces available in your Cosmic CMS environment.
Get a full list of structured content types defined in your Cosmic schema.
Review deep arrays of uploaded media to manage image limits and assets.
Ask AI about this MCP
Supported MCP Clients
Gemini Waiting for input…
Cosmic MCP Server: 10 Tools for Content & Media Ops
These tools let your agent interact with your Cosmic CMS, allowing you to manage content objects, audit schemas, and handle media assets entirely through natural language chat.
019d757dcreate cms object
Creates a new CMS object draft by accepting a structured JSON payload.
019d757dget media asset
Retrieves the specific structure and metadata for a given media ID.
019d757dget schema details
Extracts the full list of properties and fields required by an active object type.
019d757dget single object
Retrieves the complete payload and metadata for a single, specified object ID.
019d757dlist bucket objects
Lists all content types and routing spaces within the headlessly managed CMS environment.
019d757dlist schema types
Enumerates all currently active content model types defined in the schema.
019d757dlist uploaded media
Inspects all uploaded media, providing metadata and tracking image usage limits.
019d757dpatch cms object
Updates an existing CMS object by patching specific metadata blocks or values using its ID.
019d757dwipe cms object
Permanently deletes an entire CMS object and all associated content from the bucket.
019d757dwipe media asset
Deletes a specific media asset from the storage bucket, freeing up space.
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, then connect any of our 4,700+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 4,700+ 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
What you can do with this MCP connector
Cosmic MCP Server - Manage Headless CMS Content
Your AI agent gets total control over your headless CMS content. You can list, create, and patch objects, audit schema types, and manage media assets directly from chat. This lets your agent perform complex content operations—like updating a product description and clearing old media—without you ever opening the Cosmic web UI.
It's total content lifecycle management, driven by natural language.
list_bucket_objects lets your agent list all content types and routing spaces in your CMS. list_schema_types gives it a full rundown of all active content model types. get_single_object fetches the complete payload and metadata for one specific CMS object ID. list_uploaded_media lets your agent inspect all uploaded media, giving it metadata and tracking image usage limits. get_media_asset retrieves the specific structure and metadata for a given media ID.
create_cms_object builds new content drafts in Cosmic by accepting a structured JSON payload. patch_cms_object updates an existing CMS object by patching specific metadata blocks or values using its ID. wipe_cms_object permanently deletes an entire CMS object and all associated content from the bucket. get_schema_details extracts the full list of properties and fields required by an active object type. wipe_media_asset deletes a specific media asset from the storage bucket, freeing up space.
How Cosmic MCP Works
- 1 Subscribe to the server and provide your Cosmic Bucket Slug and API Write Key.
- 2 Your AI client determines the necessary action (e.g., 'Update the product page').
- 3 The server runs the specific tool (e.g.,
patch_cms_object) using the provided parameters and returns the modified object data.
The bottom line is, your agent handles the entire content workflow—from auditing the schema to publishing the final object—without you ever writing code or navigating a web UI.
Who Is Cosmic MCP For?
Content Editors who need to publish or update content without touching the web UI. Backend Developers who need to programmatically verify schema structures and field types. Product Managers who must monitor content versioning or audit media assets across multiple buckets. Growth Engineers who need to prototype content models and test updates using natural language.
Updates a product's title and metadata blocks by calling patch_cms_object or creating new drafts with create_cms_object.
Uses get_schema_details to pull field requirements for a new data model, then uses get_single_object to test object retrieval logic.
Runs list_schema_types to verify all active content models exist, then uses list_bucket_objects to monitor versioning across different content spaces.
Prototypes a new content model by running list_schema_types and then creating a test object using create_cms_object.
What Changes When You Connect
- Schema Verification: Don't guess what fields an object needs. Use
get_schema_detailsto programmatically audit an object type's properties before you even try to create it, ensuring your JSON payloads are always valid. - Atomic Content Updates: Instead of making multiple API calls, use
patch_cms_objectto update specific fields (like just the title or a metadata block) on an existing object ID. It’s surgical and safe. - Full Lifecycle Control: You can handle content from start to finish. Use
create_cms_objectto draft content,list_bucket_objectsto find the right section, and thenwipe_cms_objectwhen the content is retired. Everything is traceable. - Media Asset Management: Stop worrying about image limits.
list_uploaded_medialets your agent check the deep arrays of uploaded media, andget_media_assetlets you grab the exact metadata for a specific image. - Context Awareness: Your agent doesn't need to remember the object ID. It can run
get_single_objectto retrieve the current content, giving it the full context it needs to make the next decision or update. - Clean Development: The
list_schema_typesandlist_bucket_objectstools let you map out your entire content model structure—seeing every available content type and routing space in one chat session.
Real-World Use Cases
The Product Update Pipeline
A product manager needs to update the description and image for a key product page. They ask their agent: 'Update product ID X: set the description to 'New features available' and replace the image with ID Y.' The agent runs patch_cms_object (for the description) and get_media_asset (to verify the new image), solving the update in two steps.
Onboarding a New Content Type
A developer needs to know the exact fields for a new 'event' content type. They run list_schema_types to see what exists, then ask the agent to use get_schema_details on 'event'. The agent returns the required fields, letting the developer write the correct JSON payload for create_cms_object.
Cleaning Up Stale Content
A content team member knows a set of old blog posts are retired. They ask the agent to 'Delete all posts in the 'archive' bucket.' The agent uses list_bucket_objects to confirm the scope, then runs wipe_cms_object on the necessary IDs, cleaning the database.
Investigating Missing Media Metadata
A frontend engineer notices an image is failing to load. They ask the agent to inspect the asset. The agent runs list_uploaded_media to check the overall count, then uses get_media_asset on the specific ID to retrieve the exact metadata and figure out the imgix endpoint.
The Tradeoffs
Trying to guess schema fields
Manually writing a JSON payload for a new object and hoping the fields like 'author_bio' or 'related_topics' exist, leading to a validation error when the object fails to save.
→
First, run get_schema_details on the desired object type. This confirms the exact fields and data types (text, number, html) you need. Then, use create_cms_object with the verified structure.
Deleting content without knowing the scope
Running a generic 'delete all' command that wipes a whole content model (wipe_cms_object) without first checking if the content is truly archived or if the wrong bucket was targeted.
→
Before wiping anything, use list_bucket_objects to identify the exact routing space and content type. Use get_single_object to inspect a sample object first. This verifies you're targeting the right data.
Treating content and media separately
Updating object metadata via one tool, and then having to manually call a separate API endpoint to change the associated image, leading to mismatched data.
→
Use patch_cms_object for metadata changes, and use get_media_asset or list_uploaded_media to manage the media side. Your agent handles the coordination, ensuring both the object and the asset are updated in the same workflow.
When It Fits, When It Doesn't
Use this server if your primary job involves managing content objects (CMS) or digital assets (media) within a headless environment. Specifically, if you need your agent to perform multi-step tasks like: 1) Reading a schema (get_schema_details), 2) Fetching content (get_single_object), and 3) Modifying it (patch_cms_object).
Don't use it if you just need to list files in a standard file system (use a file storage API instead). Also, don't use it if you only need basic CRUD on a simple database table without complex metadata or media handling; a basic database connector will be simpler. This server is built for complex, structured content models.
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 server provides 10 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Manually updating content models is a nightmare of clicks and copy-pasting.
Today, updating a product description or creating a new blog post means navigating the web UI. You click 'Edit,' find the correct metadata field, copy the text, paste it, and then hit 'Save.' If you need to update five fields across three different content types, you repeat that whole sequence five times, losing context and wasting time.
With the Cosmic MCP Server, you just tell your agent: 'Update the product page for ID 123, changing the description and updating the author.' The agent executes `patch_cms_object` and handles the payload structure. You get the object updated instantly, with zero clicks and full audit logging.
Cosmic MCP Server: Manage content objects and media assets
You used to have to manually check the schema documentation to see what fields an object could accept, and then run separate commands to handle media uploads. This meant keeping track of multiple endpoints and formats.
Now, the agent uses `get_schema_details` to read the rules for you. It handles both the structured data updates (`patch_cms_object`) and the asset management (`get_media_asset`) in one flow. It's a single source of truth for your entire content stack.
Common Questions About Cosmic MCP
How do I use the `create_cms_object` tool? +
You provide the tool with a JSON payload that contains the object's title and all necessary metadata fields. The server validates this structure against the target object type before creating the draft.
Is `patch_cms_object` safe for updates? +
Yes. It only changes the specific metadata blocks you point it to, leaving all other parts of the object untouched. You must supply the object's unique ID to ensure it targets the right record.
What is the difference between `list_bucket_objects` and `list_schema_types`? +
list_schema_types shows you all the possible content structures (e.g., 'product', 'blog'). list_bucket_objects shows you all the sections or routing spaces (e.g., 'homepage', 'archive') that hold those structured objects.
How can I check the schema before writing content? +
Run get_schema_details and specify the object type name. The tool returns a structural breakdown of all required fields, data types, and validation rules.
When should I use `wipe_cms_object` vs. `wipe_media_asset`? +
Use wipe_cms_object when you need to delete an entire content record. This function removes the object and its associated metadata. wipe_media_asset only handles media files, so you use it when you need to remove a picture or asset from storage.
How do I get the structure of all available content types using `list_schema_types`? +
The list_schema_types tool returns a list of all defined content models (type slugs). You can then pass a specific type slug to get_schema_details to pull the detailed property structure for validation.
What information does `get_media_asset` return about a file? +
It returns the full structural details for a media asset, including its unique Media ID. This allows your agent to verify the file's existence and properties before use in a new object.
Can I list all content objects across different content models using `list_bucket_objects`? +
Yes, list_bucket_objects identifies all bounded routing spaces within your Cosmic environment. It lists objects regardless of their specific content type, giving you a broad overview of your content.
Can my agent create new content objects in Cosmic? +
Yes. Use the 'create_cms_object' tool. Provide the type slug, title, and a JSON object for metadata. The agent will command the backend to generate a new entry in your specified bucket.
How do I audit the schema of an existing Object Type? +
Provide the type slug to the 'get_schema_details' tool. Your agent will analyze the specific localized variables and decode the native boundaries, showing you exactly which fields and metadata parameters the schema requires.
Can I permanently delete media assets via the agent? +
Absolutely. The 'wipe_media_asset' tool allows you to irreversibly remove physical media from your Cosmic buckets. This is a powerful way to manage your storage quotas and cleanup dangling files directly from the chat.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
Deterministic JWT Inspector
Transform your AI into a cybersecurity diagnostic tool. Instantly decode and inspect JSON Web Tokens (Headers, Payloads, and Expiry metadata) without requiring signature keys.
Replicate Extended
Automate machine learning workflows via Replicate — run models, manage predictions, and search for AI assets directly from any AI agent.
Mux
Manage video assets and live streams via Mux — create assets and track performance directly from your AI agent.
You might also like
SMOTE Oversampling Engine
Balance skewed datasets instantly by generating mathematically sound synthetic minority data points via KNN.
Infisical
Manage secrets infrastructure via AI — list, create, update, and audit secrets across environments with end-to-end encryption.
Akeneo PIM
Product Information Management orchestration — manage products, attributes, and categories via AI.