Cosmos DB Container MCP. Give your agent safe, structured data access.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Azure Cosmos DB Container MCP gives your AI client secure access to one specific NoSQL document container. It lets your agent run structured queries, create new records, fetch existing documents by ID, and delete data—all without needing global cloud permissions.
This is highly scoped database management for state storage or chat histories.
What your AI agents can do
Create document
This tool lets your agent insert a new document into the container. You must provide the required ID and Partition Key details.
Delete document
The agent can permanently remove a specific document from the container, provided you supply the partition key.
Get document
Use this tool to fetch one exact document by its ID. You'll need the partition key if the container requires it.
Run specific SQL queries against all documents in the single configured container.
Create and insert brand-new documents into the database, provided you include the necessary ID and Partition Key properties.
Retrieve an exact record when you know its unique identifier and partition key.
Permanently delete documents from the container, requiring you to specify the partition key for targeting.
Ask AI about this MCP
Supported MCP Clients
OAuth 2.0 Compatible Gemini Waiting for input…
Azure Cosmos DB Container: 4 Tools
These four tools give your agent full control over the lifecycle of documents within a single NoSQL container.
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 Azure Cosmos DB Container on Vinkius019e3869create document
This tool lets your agent insert a new document into the container. You must provide the required ID and Partition Key details.
019e3869delete document
The agent can permanently remove a specific document from the container, provided you supply the partition key.
019e3869get document
Use this tool to fetch one exact document by its ID. You'll need the partition key if the container requires it.
019e3869query documents
Execute a custom SQL query against the container, allowing retrieval of multiple documents based on criteria like status or date range.
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 Azure Cosmos DB Container, then connect any of our 4,800+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 4,800+ 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 Azure Cosmos DB Container. 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 4 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Managing structured data means constant context switching.
Right now, if your agent needs to read a user's profile and then update their status, you have to write complex code that handles fetching the initial document by ID, parsing the JSON body in memory, modifying the required fields, and finally sending a separate patch request. It’s slow, error-prone boilerplate.
With this MCP, your agent simply asks for what it needs. The system handles the full lifecycle: retrieving the data via `get_document`, letting the AI adjust the payload, and then saving the complete change using `create_document`. You just get the result—the state is updated.
The Cosmos DB Container MCP provides atomic document control.
You no longer have to rely on brittle, multi-step API calls for basic data operations. The dedicated tools mean you can isolate read actions using `get_document` and write actions using `create_document`, making the code much cleaner.
It's direct. You call a function, it talks to Azure Cosmos DB, and it returns. No more middleware needed just to handle simple data persistence.
What you can do with this MCP connector
This connector strips away all the dangerous global Azure permissions. What it gives your AI client is one surgical capability: full read/write access to a single Cosmos DB container. You can safely manage structured data, whether you're storing user profiles, session tokens, or complex chat threads. It handles rich NoSQL queries and keeps everything contained within that single boundary.
Because the scope is so tight, your agent only sees what it needs, preventing accidental exposure to critical databases. If you're building an application that requires reliable state management but doesn't need full cloud admin privileges, this MCP makes it simple. Vinkius hosts this connection, letting you plug secure document storage into any AI client.
019e3869-b80d-7286-a03c-0004c3c6fe7f 
How Cosmos DB Container MCP Works
- 1 Your agent determines it needs data and uses
query_documentsto execute a targeted SQL query. - 2 The MCP connects directly to Cosmos DB, runs the query against the single container, and returns the result set to your agent.
- 3 Your agent processes the returned documents, allowing it to make decisions or pass the data back into its workflow.
The bottom line is that your AI client gets a safe, focused pipe directly into structured NoSQL data without needing any elevated cloud access rights.
Who Is Cosmos DB Container MCP For?
System architects and backend engineers who need to give their agents stateful memory or limited persistence. This is for the developer tired of writing complex database connection wrappers just to store chat logs.
Needs a reliable place to save user session data, ensuring the agent can recall context across multiple interactions.
Builds microservices that need simple document persistence for configuration settings or transient state tracking.
Designs agent workflows that require the ability to save and retrieve specific, structured knowledge documents.
What Changes When You Connect
- Security is contained. Because the agent's scope is locked to one container, you eliminate the risk of global cloud misconfigurations or accidental data deletion in other areas.
- Handle state easily. You can use
create_documentandget_documentto save conversation history or application states reliably so your agent never 'forgets' context. - Complex searches are simple. Instead of building custom search APIs, you just run a query with
query_documents, letting the AI execute complex SQL logic directly. - It works out of the box. This MCP instantly gives your system a scalable NoSQL database layer for structured memories and application state without any setup headaches.
- Clean data lifecycle management. You can use
delete_documentwhen specific records are obsolete, keeping your container clean and accurate.
Real-World Use Cases
Tracking user feedback tickets
A support agent's workflow needs to update the status of a ticket. The agent uses get_document to fetch the current record, modifies the 'status' field (e.g., from 'open' to 'assigned'), and then calls create_document to save the updated state.
Building an inventory tracker
The system needs to know which items are low on stock. The agent runs a query using query_documents with criteria like 'quantity < 10'. This instantly returns all necessary SKUs for the ops team.
Archiving old chat logs
When a conversation hits its retention limit, the agent uses delete_document, specifying the unique ID and partition key, to safely archive or remove the data point from the active container.
The Tradeoffs
Trying to query all data at once
Asking the agent to 'just dump everything in the database.' This is inefficient and often hits rate limits.
→
Always use query_documents. Filter your request by specific criteria, like a date range or status field (e.g., SELECT * FROM c WHERE c.status = @status). Don't query everything.
Overwriting data without checking
Using create_document blindly to save a profile, potentially overwriting recent user changes by forgetting the current ID.
→
First, use get_document to retrieve the existing record. Then, pass both the old data and the new updates into your process before saving.
Assuming global access
Writing code that needs full read/write across multiple databases in Azure.
→ Stick to this MCP. It's scoped only to one container, giving you the necessary data without needing dangerous global cloud permissions.
When It Fits, When It Doesn't
Use this MCP if your persistence layer is a single, structured NoSQL document store and you need fine-grained control over CRUD operations (e.g., managing chat logs or application state). You must be comfortable defining the schema for that contained container. Don't use it if: 1) Your data relationships are complex graphs; you need a graph database tool instead. 2) You require multi-tenant isolation across many different containers; this MCP is locked to one scope. If your needs involve transactional integrity (ACID guarantees), check out relational database connectors.
Common Questions About Cosmos DB Container MCP
How do I query documents using the `query_documents` tool? +
You execute a standard SQL query against the container. Remember, you don't include 'DB' or 'Container' names in the SELECT statement; just focus on the structure: SELECT * FROM c WHERE c.status = @status.
Can I use `create_document` if I already have an ID? +
Yes, you must provide both the unique ID and the Partition Key within the document payload when calling create_document. This ensures the record lands exactly where it should.
What is required for the `delete_document` tool? +
You must specify the partition key. The agent needs this specific piece of metadata to correctly target and delete a document without risk.
Does `get_document` only retrieve one record? +
Yes, get_document retrieves exactly one document based on its unique ID. If you need multiple records matching criteria, use query_documents instead.
When I use `create_document`, do I always need to specify both an ID and a Partition Key? +
Yes, in almost all scenarios, you must provide both a unique document ID and the appropriate partition key value. This ensures data is placed correctly within the container structure.
What specific syntax do I need for the `query_documents` tool to run successfully? +
You need to write standard SQL queries, following the pattern like "SELECT * FROM c WHERE...". Importantly, never include the name of the database or container in your query string.
If I use `get_document` and the specified ID does not exist, what response should my agent expect? +
The tool will return an explicit error message or a null result. Your agent should check for this failure state to handle missing records gracefully.
Does using this MCP give my AI client permission to access databases outside of the configured container? +
No, absolutely not. The MCP is strictly scoped and locked down to manage only documents within your single specified Cosmos DB Container. It cannot affect other cloud resources.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.