Azure AI Search Vector Store

Azure AI Search (formerly Azure Cognitive Search) is a cloud search service with vector search capabilities for RAG and semantic search applications. Use this node to store, retrieve, and query vector embeddings alongside their content and metadata.

On this page, you'll find the node parameters for the Azure AI Search Vector Store node, and links to more resources.

Prerequisites

Before using this node, you need:

1. An Azure subscription
2. An Azure AI Search service
3. API key authentication configured (admin key for write operations, query key for read-only)

See credentials documentation for setup instructions.

Index configuration

The node automatically creates indexes if they don't exist. When auto-creating, the node configures:

• Vector fields with appropriate dimensions based on your embeddings model
• HNSW algorithm for efficient similarity search with cosine metric
• Content and metadata fields for filtering and retrieval

You can also pre-create indexes in Azure Portal for custom configurations. Example schema:

{
  "name": "n8n-vectorstore",
  "fields": [
    {
      "name": "id",
      "type": "Edm.String",
      "key": true,
      "filterable": true
    },
    {
      "name": "content",
      "type": "Edm.String",
      "searchable": true
    },
    {
      "name": "content_vector",
      "type": "Collection(Edm.Single)",
      "searchable": true,
      "vectorSearchDimensions": 1536,
      "vectorSearchProfileName": "n8n-vector-profile"
    },
    {
      "name": "metadata",
      "type": "Edm.String",
      "filterable": true
    }
  ],
  "vectorSearch": {
    "profiles": [
      {
        "name": "n8n-vector-profile",
        "algorithm": "n8n-vector-algorithm"
      }
    ],
    "algorithms": [
      {
        "name": "n8n-vector-algorithm",
        "kind": "hnsw",
        "hnswParameters": {
          "metric": "cosine",
          "m": 4,
          "efConstruction": 400,
          "efSearch": 500
        }
      }
    ]
  }
}

Node usage patterns

Use as a regular node to insert and retrieve documents

Use the node directly in workflows to insert or retrieve documents without an agent. See this template for an example pattern (uses Supabase, but the pattern is identical).

Connect directly to an AI agent as a tool

Connect to an AI agent's tool connector to use the vector store as a searchable knowledge base:

AI agent (tools connector) → Azure AI Search Vector Store node

Use a retriever to fetch documents

Use with Vector Store Retriever and Question and Answer Chain for retrieval-augmented generation:

Question and Answer Chain (Retriever) → Vector Store Retriever (Vector Store) → Azure AI Search Vector Store

See this example workflow.

Use the Vector Store Question Answer Tool

Use Vector Store Question Answer Tool to summarize and answer questions:

AI agent (tools) → Vector Store Question Answer Tool (Vector Store) → Azure AI Search Vector Store

See this example.

Node parameters

This Vector Store node has five modes: Get Many, Insert Documents, Retrieve Documents (As Vector Store for Chain/Tool), Retrieve Documents (As Tool for AI Agent), and Update Documents. The mode you select determines the operations you can perform with the node and what inputs and outputs are available.

Get Many

In this mode, you can retrieve multiple documents from your vector database by providing a prompt. The prompt will be embedded and used for similarity search. The node will return the documents that are most similar to the prompt with their similarity score. This is useful if you want to retrieve a list of similar documents and pass them to an agent as additional context.

Insert Documents

Use Insert Documents mode to insert new documents into your vector database.

Retrieve Documents (As Vector Store for Chain/Tool)

Use Retrieve Documents (As Vector Store for Chain/Tool) mode with a vector-store retriever to retrieve documents from a vector database and provide them to the retriever connected to a chain. In this mode you must connect the node to a retriever node or root node.

Retrieve Documents (As Tool for AI Agent)

Use Retrieve Documents (As Tool for AI Agent) mode to use the vector store as a tool resource when answering queries. When formulating responses, the agent uses the vector store when the vector store name and description match the question details.

Update Documents

Use Update Documents mode to update documents in a vector database by ID. Fill in the ID with the ID of the embedding entry to update.

Rerank Results

Enables reranking. If you enable this option, you must connect a reranking node to the vector store. That node will then rerank the results for queries. You can use this option with the Get Many, Retrieve Documents (As Vector Store for Chain/Tool) and Retrieve Documents (As Tool for AI Agent) modes.

Get Many parameters

• Endpoint: Your Azure AI Search endpoint (format: https://your-service.search.windows.net)
• Index Name: The index to query
• Limit: Maximum documents to return (default: 4)

Insert Documents parameters

• Endpoint: Your Azure AI Search endpoint
• Index Name: The index to use (created automatically if it doesn't exist)
• Batch Size: Number of documents uploaded per batch to Azure AI Search. Adjust based on document size and your service tier limits. This controls upload batching only—embedding generation batching is configured in embedding nodes.

Update Documents parameters

• Endpoint: Your Azure AI Search endpoint
• Index Name: The index to update

Retrieve Documents parameters (As Vector Store for Chain/Tool)

• Endpoint: Your Azure AI Search endpoint
• Index Name: The index to query

Retrieve Documents (As Tool for AI Agent) parameters

• Name: Tool name shown to the LLM
• Description: Explain to the LLM what this tool does. Be specific to help the LLM choose when to use this tool.
• Endpoint: Your Azure AI Search endpoint
• Index Name: The index to query
• Limit: Maximum results to retrieve (e.g., 10 for ten best matches)

Node options

Options

• Filter: OData filter expression to filter results by document fields or metadata. See filter examples below.
• Query Mode: Search strategy to use:
• Vector: Similarity search using embeddings only
• Keyword: Full-text search using BM25 ranking
• Hybrid (default): Combines vector and keyword search with Reciprocal Rank Fusion (RRF)
• Semantic Hybrid: Hybrid search with semantic reranking for improved relevance
• Semantic Configuration: Name of the semantic configuration to use for semantic ranking. Defaults to semantic-search-config if not specified. Only required if you pre-created an index with a custom semantic configuration name.

OData filter examples

Azure AI Search uses OData syntax for filtering. Metadata fields are accessed using metadata/fieldName format.

Filter by document ID:

id eq '3da6491a-f930-4a4e-9471-c05dcd450ba0'

Filter by metadata field:

metadata/source eq 'user-guide'

Complex AND filter:

metadata/category eq 'technology' and metadata/author eq 'John'

Complex OR filter:

metadata/source eq 'user-guide' or metadata/rating ge 4

Numeric comparison:

metadata/rating ge 4 and metadata/rating lt 10

String matching with NOT:

metadata/category eq 'technology' and metadata/title ne 'Deprecated'

Supported OData operators:

• Comparison: eq, ne, gt, ge, lt, le
• Logical: and, or, not
• String functions: startswith(), endswith(), contains()
• Collection functions: any(), all()

Azure AI Search specific features

Hybrid search with RRF

Azure AI Search's hybrid search uses Reciprocal Rank Fusion to merge vector and keyword results, providing better accuracy than either method alone.

Semantic ranking

Semantic Hybrid mode applies machine learning models to rerank results based on semantic understanding of your query. This requires a semantic configuration in your index.

OData filters

Use OData syntax to filter by document fields or metadata before vector search executes. This improves performance and precision when you need results from specific sources or with certain attributes.

HNSW algorithm

Azure AI Search uses Hierarchical Navigable Small World (HNSW) graphs for approximate nearest neighbor search, providing fast retrieval at scale with configurable accuracy/speed tradeoffs.

Troubleshooting

Index issues

Index not found: Verify the index name is correct (case-sensitive) and exists in your Azure AI Search service. If using auto-creation, check that the index was created successfully.

Vector dimension mismatch: Ensure your embedding model dimensions match the index vector field dimensions. Check the index schema to confirm the vectorSearchDimensions setting.

Document insert failures:

• Verify write permissions (admin API key required)
• Check document fields match your index schema
• Ensure required fields are provided in documents
• Review batch size settings if experiencing timeouts with large document sets

Filter issues

Filter not working:

• Verify OData syntax is correct
• Ensure metadata fields use metadata/ prefix: metadata/source eq 'value'
• Check that filtered fields are marked as filterable in your index schema
• Test with simple filters first (id eq 'value') before complex expressions

Invalid OData syntax:

• Use single quotes for string values: metadata/source eq 'value'
• Use proper operators: eq, ne, gt, ge, lt, le, and, or, not
• Refer to OData filter documentation for syntax details

Connection issues

Unable to connect:

• Verify endpoint URL format: https://your-service.search.windows.net
• Confirm your Azure AI Search service is running and accessible
• Check network security groups, firewall rules, and private endpoint configurations
• For Azure-hosted n8n, verify virtual network peering or service endpoint configuration if using private endpoints

Authentication issues

For authentication troubleshooting including API key errors, refer to the credentials documentation troubleshooting section.

Templates and examples

Browse Azure AI Search Vector Store node documentation integration templates or search all templates

Related resources

• Azure AI Search Vector Search documentation
• LangChain Azure AI Search integration
• Azure AI Search REST API reference
• OData filter syntax for Azure AI Search

View n8n's Advanced AI documentation.

New to working with AI and using self-hosted n8n? Try n8n's self-hosted AI Starter Kit to get started with a proof-of-concept or demo playground using Ollama, Qdrant, and PostgreSQL.