n8n-nodes-falkordb

v1.0.0

Published

6 months ago

n8n node for interacting with FalkorDB graph database

0High
0Medium
0Low

paerrin

n8n-community-node-package n8n falkordb graph-database cypher knowledge-graph

@falkordb/n8n-nodes-falkordb

This is an n8n community node that lets you interact with FalkorDB graph database in your n8n workflows.

FalkorDB is a super fast graph database that uses GraphBLAS for its sparse adjacency matrix graph representation, optimized for Knowledge Graphs and GraphRAG applications.

n8n is a fair-code licensed workflow automation platform.

Installation

Follow the installation guide in the n8n community nodes documentation.

Community Nodes (Recommended)

Go to Settings > Community Nodes
Select Install
Enter @falkordb/n8n-nodes-falkordb in Enter npm package name
Select Install

Manual Installation

npm install @falkordb/n8n-nodes-falkordb

Configuration

Credentials

To connect to FalkorDB, you'll need:

Host: FalkorDB server hostname or IP (default: localhost)
Port: FalkorDB server port (default: 6379)
Username: FalkorDB username (default: default)
Password: FalkorDB password (leave empty for default user)
Base URL: FalkorDB REST API base URL (e.g., http://localhost:3000)
Use TLS: Enable if using TLS/SSL connection

Operations

Graph

List: Get all available graphs
Count: Count nodes and relationships in a graph
Create: Create a new graph
Delete: Delete a graph
Duplicate: Duplicate/copy a graph
Export: Export graph data
Get Info: Get graph metadata (labels, relationships, properties, functions)
Rename: Rename a graph

Index

List Indexes: List all indexes in a graph
Create Full-Text Index: Create a full-text search index for text search
Create Vector Index: Create a vector similarity index for semantic search
Vector Search: Search using vector similarity (perfect for RAG applications)

Query

Execute: Execute a Cypher query
Execute Batch: Execute multiple queries in batch (one per line)
Explain: Get query execution plan without running it
Profile: Get query performance metrics

Node

Get: Retrieve node details
Delete: Delete a node or relationship
Add Label: Add a label to a node
Remove Label: Remove a label from a node
Set Property: Set a property on a node

Complete RAG Example (Ollama + FalkorDB)

Setup: Index Your Documents

1. HTTP Request - Read documents from your source
2. For each document:
   a. Ollama - Generate Embedding
      - Model: nomic-embed-text
      - Prompt: {{$json.content}}
   b. FalkorDB - Execute Query
      - Graph: knowledge_base
      - Query: CREATE (d:Document {
                 id: '{{$json.id}}',
                 content: '{{$json.content}}',
                 embedding: {{$json.embedding}}
               })
3. FalkorDB - Create Vector Index
   - Graph: knowledge_base
   - Index Name: doc_embeddings
   - Label: Document
   - Property: embedding
   - Dimensions: 768

Query: Semantic Search + Answer

1. Ollama - Generate Embedding
   - Model: nomic-embed-text
   - Input: User question
2. FalkorDB - Vector Search
   - Graph: knowledge_base
   - Search Vector: {{$json.embedding}}
   - Top K: 3
3. Code - Combine context
   - Extract document contents from top 3 results
4. Ollama - Chat Completion
   - Model: llama3.2
   - System: You are a helpful assistant
   - Context: {{$json.combined_context}}
   - Question: {{$input.question}}

Example Workflows

Create a Knowledge Graph (with Batch)

1. FalkorDB - Create Graph (graphName: "knowledge_base")
2. FalkorDB - Execute Batch
   - Graph: knowledge_base
   - Queries:
     CREATE (p:Person {name: 'Alice', age: 30})
     CREATE (c:Company {name: 'TechCorp'})
     MATCH (p:Person {name: 'Alice'}), (c:Company {name: 'TechCorp'})
     CREATE (p)-[:WORKS_AT]->(c)

Vector Search RAG Workflow (Ollama)

1. Ollama - Generate Embedding
   - Model: nomic-embed-text (768 dimensions)
   - Input: User's search query
   - Output: Embedding vector
2. FalkorDB - Vector Search
   - Graph: documents
   - Search Vector: {{$json.embedding}}
   - Top K: 5
3. Process top 5 similar documents
4. Ollama - Generate response using retrieved context
   - Model: llama3.2 or mistral

Alternative: Open WebUI Embeddings

1. HTTP Request - POST to Open WebUI
   - URL: http://localhost:8080/api/v1/embeddings
   - Body: {"model": "nomic-embed-text", "input": "{{$json.query}}"}
2. FalkorDB - Vector Search
   - Search Vector: {{$json.data[0].embedding}}

Create Vector Index for Semantic Search

1. FalkorDB - Create Vector Index
   - Graph: documents
   - Index Name: doc_embeddings
   - Label: Document
   - Property: embedding
   - Vector Dimensions: 768 (nomic-embed-text) or 1024 (mxbai-embed-large)
2. Ready to perform vector similarity searches!

Full-Text Search

1. FalkorDB - Create Full-Text Index
   - Graph: articles
   - Index Name: article_content
   - Label: Article
   - Property: content
2. FalkorDB - Execute Query
   - Query: CALL db.idx.fulltext.queryNodes('article_content', 'search terms')

Export and Duplicate Graph

1. FalkorDB - Export
   - Graph: production_graph
2. FalkorDB - Duplicate
   - Source Graph: production_graph
   - New Graph Name: backup_graph
3. Backup complete!

Embedding Model Dimensions

Common embedding models and their vector dimensions:

Ollama Models

nomic-embed-text: 768 dimensions (recommended, fast & accurate)
mxbai-embed-large: 1024 dimensions (high quality)
all-minilm: 384 dimensions (lightweight)

OpenAI Models

text-embedding-3-small: 1536 dimensions
text-embedding-3-large: 3072 dimensions
text-embedding-ada-002: 1536 dimensions (legacy)

Open Source Models

sentence-transformers/all-MiniLM-L6-v2: 384 dimensions
BAAI/bge-small-en-v1.5: 384 dimensions
BAAI/bge-base-en-v1.5: 768 dimensions
BAAI/bge-large-en-v1.5: 1024 dimensions

Important: The vector dimensions in your index must match your embedding model!

Compatibility

n8n version: 1.0.0+
FalkorDB version: 4.0+
REST API: Requires FalkorDB REST API endpoint
Embedding Providers: Ollama, Open WebUI, OpenAI, HuggingFace, or any custom embedding service

Resources

Support

For issues or questions:

GitHub Issues: n8n-nodes-falkordb/issues
FalkorDB Community: FalkorDB Discussions

License

MIT