[FEATURE] Add Azure DocumentDB vector search tool integration

Feature Area

Integration with external tools

Is your feature request related to a an existing bug? Please link it here.

CrewAI currently has a MongoDBVectorSearchTool, but there is no dedicated tool integration for Azure DocumentDB.

Azure DocumentDB is a fully managed NoSQL database service with MongoDB compatibility and native vector search capabilities. Users building RAG agents on Azure DocumentDB currently need to write custom retrieval code instead of using a first-class CrewAI tool.

A dedicated AzureDocumentDBVectorSearchTool would make it easier to use Azure DocumentDB as a vector store for CrewAI agents while keeping the integration clearly scoped to Azure DocumentDB configuration, documentation, authentication, indexing, and vector search behavior.

Describe the solution you'd like

Add a new Azure DocumentDB vector search tool to crewai-tools, separate from the existing MongoDBVectorSearchTool.

Proposed public API:

• AzureDocumentDBVectorSearchTool
• AzureDocumentDBVectorSearchConfig
• AzureDocumentDBToolSchema

The tool should support:

• Connecting to Azure DocumentDB using a connection string
• Selecting database and collection
• Generating embeddings using the existing CrewAI/OpenAI or Azure OpenAI embedding pattern
• Adding text documents with metadata and embeddings
• Creating or documenting the required Azure DocumentDB vector index configuration
• Running vector similarity search against Azure DocumentDB
• Returning JSON-serializable search results for CrewAI agents
• Optional query configuration such as limit, filters, score output, and whether to include embeddings

This should be implemented as a new tool package, for example:

crewai_tools/tools/azure_documentdb_vector_search_tool/

The README should include setup instructions, required Azure DocumentDB resource configuration, environment variables, ingestion example, and CrewAI Agent usage example.

Reference Azure DocumentDB docs:
https://learn.microsoft.com/en-us/azure/documentdb/

Describe alternatives you've considered

An alternative is to use the existing MongoDBVectorSearchTool because Azure DocumentDB has MongoDB compatibility. However, that tool is branded and documented for MongoDB/Atlas and uses MongoDB-specific assumptions.

Another alternative is for users to write custom retrieval code outside CrewAI, but that creates duplicated boilerplate for connection handling, embeddings, indexing guidance, ingestion, and query execution.

A dedicated Azure DocumentDB tool would provide a clearer user experience for Azure users and allow Azure-specific documentation, examples, configuration, and future enhancements without overloading the MongoDB integration.

Additional context

Azure DocumentDB documentation describes vector search and embeddings support:
https://learn.microsoft.com/en-us/azure/documentdb/

Related Azure DocumentDB vector search quickstarts:

• https://learn.microsoft.com/en-us/azure/documentdb/quickstart-python-vector-search
• https://learn.microsoft.com/en-us/azure/documentdb/quickstart-nodejs-vector-search
• https://learn.microsoft.com/en-us/azure/documentdb/quickstart-dotnet-vector-search

Existing CrewAI tool for reference:
https://github.com/crewAIInc/crewAI/tree/main/lib/crewai-tools/src/crewai_tools/tools/mongodb_vector_search_tool

The goal is not to modify the MongoDB extension, but to add a new Azure DocumentDB integration with its own name, README, examples, optional dependency, and tests.

Willingness to Contribute

Yes, I'd be happy to submit a pull request

Read the existing crewai_tools/tools/mongodb_vector_search_tool/ package and the linked Azure DocumentDB vector-search quickstarts first. Define the new package around the proposed public API, Azure connection and vector-search requirements, then add its README, examples, optional dependency, and tests. Done means the integration supports ingestion and JSON-serializable search results without modifying the MongoDB tool.