n8n-nodes-elasticsearch-vectorstore

Production-ready n8n community node that provides an Elasticsearch Vector Store Tool node with upsert and similarity search operations. Supports Elasticsearch 7.5 and 8.11, using the built-in Elasticsearch credential and Embedding node input.

Features

• Upsert documents with embeddings into Elasticsearch
• Similarity search using kNN (ES 8.11) or script_score fallback (ES 7.5)
• Embedding node input (ai_embedding)
• Optional index creation with mapping
• Deterministic hash IDs based on text+metadata

Compatibility

• Elasticsearch 7.5 and 8.11
• n8n self-hosted (including Kubernetes)

Credentials

• Elasticsearch API: Uses the same fields as n8n built-in Elasticsearch credential (basic auth, base URL, SSL ignore)

Node Parameters

• Connection: Elasticsearch credentials, Index Name
• Embeddings: Embedding node input, vector dimensions, batch size
• Upsert: text/id/metadata fields, auto-generate ID, ensure index
• Search: search mode (AUTO/KNN/SCRIPT_SCORE), topK, minScore

Build & Run

npm install
npm run build
npm pack

Install in n8n

Option A: Community package installation (UI)

1. Enable community packages in n8n (Settings → Community Nodes).
2. Install n8n-nodes-elasticsearch-vectorstore.

Option B: Bake into a custom Docker image (K8s friendly)

FROM n8nio/n8n:latest
USER root
RUN npm install /path/to/n8n-nodes-elasticsearch-vectorstore.tgz
USER node

Restart the deployment after updating the image.

Option C: Mount as custom extension

Build and copy the package to the custom extensions directory:

export N8N_CUSTOM_EXTENSIONS=/home/node/.n8n/custom
mkdir -p /home/node/.n8n/custom
cp -R /path/to/n8n-nodes-elasticsearch-vectorstore /home/node/.n8n/custom

Restart n8n.

How to test locally

npm install
npm run dev

Open http://localhost:5678, create a workflow, and add the Elasticsearch Vector Store Tool node.

Using an Embedding Node Input

Set Embedding Source to Use Embedding Node and connect your Embeddings node to the Embeddings input of the Elasticsearch Vector Store Tool node.

Use as AI Agent Tool

Set Operation to Similarity Search (As Tool) and connect the Elasticsearch Vector Store Tool node to the AI Agent Tool port. Connect an Embeddings node to the Embeddings input.

Example workflow JSON (Upsert)

{
  "nodes": [
    {
      "parameters": {
        "operation": "upsert",
        "indexName": "documents",
        "ensureIndex": true,
        "vectorDimensions": 1536,
        "batchSize": 10,
        "textField": "text",
        "idField": "id",
        "metadataField": "metadata",
        "autoId": "hash"
      },
      "name": "Elasticsearch Vector Store Tool",
      "type": "n8n-nodes-elasticsearch-vectorstore.elasticsearchVectorStoreTool",
      "typeVersion": 1,
      "position": [520, 300],
      "credentials": {
        "elasticsearchApi": {
          "id": "1",
          "name": "ES Cluster"
        }
      }
    }
  ],
  "connections": {}
}

Example workflow JSON (Similarity Search)

{
  "nodes": [
    {
      "parameters": {
        "operation": "search",
        "indexName": "documents",
        "ensureIndex": false,
        "vectorDimensions": 1536,
        "batchSize": 5,
        "query": "Find related documents to prompt engineering",
        "searchMode": "auto",
        "topK": 5,
        "minScore": 0
      },
      "name": "Elasticsearch Vector Store Tool",
      "type": "n8n-nodes-elasticsearch-vectorstore.elasticsearchVectorStoreTool",
      "typeVersion": 1,
      "position": [520, 300],
      "credentials": {
        "elasticsearchApi": {
          "id": "1",
          "name": "ES Cluster"
        }
      }
    }
  ],
  "connections": {}
}