Knowledge Graph Builder

Transform unstructured data (PDFs, DOCs, TXTs, YouTube videos, web pages, etc.) into a structured Knowledge Graph stored in Neo4j using the power of Large Language Models (LLMs) and the LangChain framework.

This application allows you to upload files from various sources (local machine, GCS, S3 bucket, or web sources), choose your preferred LLM model, and generate a Knowledge Graph.

Getting Started

Prerequisites

• Python 3.12 or higher (for local/separate backend deployment)
• Neo4j Database 5.23 or later with APOC installed.
  • Neo4j 5.23 is required because the backend uses the Cypher variable-scope subquery syntax (CALL (variable) { ... }), which is not supported by earlier Neo4j 5.x releases such as 5.20.
  • Neo4j Aura databases (including the free tier) are supported.
  • If using Neo4j Desktop, you will need to deploy the backend and frontend separately (docker-compose is not supported).

Backend Setup

1. Create a .env file in the backend folder by copying backend/example.env.
2. Pre-configure user credentials in the .env file to bypass the login dialog:

   NEO4J_URI=<your-neo4j-uri>
   NEO4J_USERNAME=<your-username>
   NEO4J_PASSWORD=<your-password>
   NEO4J_DATABASE=<your-database-name>
   

3. Run:

   cd backend
   python3.12 -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   pip install -r requirements.txt
   uvicorn score:app --reload
   

Key Features

Knowledge Graph Creation

• Seamlessly transform unstructured data into structured Knowledge Graphs using advanced LLMs.
• Extract nodes, relationships, and their properties to create structured graphs.

Schema Support

• Use a custom schema or existing schemas configured in the settings to generate graphs.

Graph Visualization

• View graphs for specific or multiple data sources simultaneously in Neo4j Bloom.

Chat with Data

• Interact with your data in the Neo4j database through conversational queries.
• Retrieve metadata about the source of responses to your queries.
• For a dedicated chat interface, use the standalone chat application with the /chat-only route.

LLMs Supported

1. OpenAI
2. Gemini
3. Diffbot
4. Azure OpenAI (dev deployed version)
5. Anthropic (dev deployed version)
6. Fireworks (dev deployed version)
7. Groq (dev deployed version)
8. Amazon Bedrock (dev deployed version)
9. Ollama (dev deployed version)
10. Deepseek (dev deployed version)
11. Other OpenAI-compatible base URL models (dev deployed version)

Token Usage Tracking

• Easily monitor and track your LLM token usage for each user and database connection.
• Enable this feature by setting the TRACK_USER_USAGE environment variable to true in your backend configuration.
• View your daily and monthly token consumption and limits, helping you manage usage and avoid overages.
• You can check your remaining token limits at any time using the provided API endpoint.

Embedding Model Selection

• Choose from a variety of embedding models to generate vector embeddings for your data. This can be configured from the frontend in Graph Settings > Processing Configuration > Select Embedding Model.
• Supported model providers include OpenAI, Gemini, Amazon Titan, and Sentence Transformers.
• Your selected embedding model is saved to your user profile when TRACK_USER_USAGE is enabled.

Local Configuration

You have two ways to configure the embedding model locally:

1. With User Tracking (TRACK_USER_USAGE=true):

   • Set TRACK_USER_USAGE to true in your backend .env file.
   • Provide your token tracking database credentials (TOKEN_TRACKER_DB_URI, TOKEN_TRACKER_DB_USERNAME, etc.).
   • Select your desired embedding model from the frontend. Your selection will be saved and automatically used in subsequent sessions.

2. Without User Tracking (TRACK_USER_USAGE=false):

   • Set TRACK_USER_USAGE to false.
   • Specify the embedding model and provider directly in your backend .env file using EMBEDDING_MODEL and EMBEDDING_PROVIDER.
   • If these variables are not set, the application defaults to a Sentence Transformer model.
   • In this mode, the embedding model cannot be changed from the frontend.

---

Getting Started

Prerequisites

• Neo4j Database 5.23 or later with APOC installed.
  • Neo4j 5.23 is required because the backend uses the Cypher variable-scope subquery syntax (CALL (variable) { ... }), which is not supported by earlier Neo4j 5.x releases such as 5.20.
  • Neo4j Aura databases (including the free tier) are supported.
  • If using Neo4j Desktop, you will need to deploy the backend and frontend separately (docker-compose is not supported).

---

Deployment Options

Local Deployment

Using Docker-Compose

Run the application using the default docker-compose configuration.

1. Supported LLM Models:
   By default, only OpenAI and Diffbot are enabled. Gemini requires additional GCP configurations.
   Use the VITE_LLM_MODELS_PROD variable to configure the models you need. Example:

   VITE_LLM_MODELS_PROD="gemini_3.5_flash,openai_gpt_5.4_mini,diffbot,anthropic_claude_4.5_haiku"
   

2. Anthropic Models: Use the latest Claude model in your config:

   LLM_MODEL_CONFIG_ANTHROPIC_CLAUDE_4_7_OPUS="claude-opus-4-7,anthropic_api_key"
   

3. Input Sources:
   By default, the following sources are enabled: local, YouTube, Wikipedia, AWS S3, and web.
   To add Google Cloud Storage (GCS) integration, include gcs and your Google client ID:

   VITE_REACT_APP_SOURCES="local,youtube,wiki,s3,gcs,web"
   VITE_GOOGLE_CLIENT_ID="your-google-client-id"
   

Chat Modes

Configure chat modes using the VITE_CHAT_MODES variable:

• By default, all modes are enabled: vector, graph_vector, graph, fulltext, graph_vector_fulltext, entity_vector, and global_vector.
• To specify specific modes, update the variable. For example:

  VITE_CHAT_MODES="vector,graph"
  

---

Running Backend and Frontend Separately

For development, you can run the backend and frontend independently.

Frontend Setup

1. Create a .env file in the frontend folder by copying frontend/example.env.
2. Update environment variables as needed.
3. Run:

   cd frontend
   

yarn yarn run dev

#### **Backend Setup**
1. Create a `.env` file in the `backend` folder by copying `backend/example.env`.
2. Pre-configure user credentials in the `.env` file to bypass the login dialog:
```bash
NEO4J_URI=<your-neo4j-uri>
NEO4J_USERNAME=<your-username>
NEO4J_PASSWORD=<your-password>
NEO4J_DATABASE=<your-database-name>

3. Run:

   cd backend
   

python -m venv envName source envName/bin/activate pip install -r requirements.txt uvicorn score:app --reload

---

### **Cloud Deployment**

Deploy the application on **Google Cloud Platform** using the following commands:

#### **Frontend Deployment**
```bash
gcloud run deploy dev-frontend \
--source . \
--region us-central1 \
--allow-unauthenticated

Backend Deployment

gcloud run deploy dev-backend \
  --set-env-vars "OPENAI_API_KEY=<your-openai-api-key>" \
  --set-env-vars "DIFFBOT_API_KEY=<your-diffbot-api-key>" \
  --set-env-vars "NEO4J_URI=<your-neo4j-uri>" \
  --set-env-vars "NEO4J_USERNAME=<your-username>" \
  --set-env-vars "NEO4J_PASSWORD=<your-password>" \
  --source . \
  --region us-central1 \
  --allow-unauthenticated

---

For local llms (Ollama)

1. Pull the docker image of ollama

   docker pull ollama/ollama
   

2. Run the ollama docker image

   docker run -d -v ollama:/root/.ollama -p 11434:11434 --name ollama ollama/ollama
   

3. Execute any llm model, e.g., llama3

   docker exec -it ollama ollama run llama3
   

4. Configure env variable in docker compose.

   LLM_MODEL_CONFIG_ollama_<model_name>
   # example
   LLM_MODEL_CONFIG_ollama_llama3=${LLM_MODEL_CONFIG_ollama_llama3-llama3,http://host.docker.internal:11434}
   

5. Configure the backend API url

   VITE_BACKEND_API_URL=${VITE_BACKEND_API_URL-backendurl}
   

6. Open the application in browser and select the ollama model for the extraction.
7. Enjoy Graph Building.

---

Usage

1. Connect to a Neo4j Aura Instance, which can be either AURA DS or AURA DB, by passing the URI and password through the backend environment, filling in the login dialog, or dragging and dropping the Neo4j credentials file.
2. To differentiate, we have added different icons. For AURA DB, there is a database icon, and for AURA DS, there is a scientific molecule icon right under the Neo4j Connection details label.
3. Choose your source from a list of unstructured sources to create a graph.
4. Change the LLM (if required) from the dropdown, which will be used to generate the graph.
5. Optionally, define the schema (nodes and relationship labels) in the entity graph extraction settings.
6. Either select multiple files to 'Generate Graph', or all the files in 'New' status will be processed for graph creation.
7. View the graph for individual files using 'View' in the grid, or select one or more files and 'Preview Graph'.
8. Ask questions related to the processed/completed sources to the chatbot. Also, get detailed information about your answers generated by the LLM.

---

ENV

Example Environment Files

Refer to the example environment files for additional variables and configuration:

• Backend example.env
• Frontend example.env

---

Cloud Build Deployment

You can deploy the backend and the frontend to Google Cloud Run using Cloud Build, either manually or via automated triggers.

Automated Deployment (Recommended)

1. Connect your repository to Google Cloud Build:

   • In the Google Cloud Console, go to Cloud Build > Triggers.
   • Create a new trigger and select your repository.
   • Set the trigger to run on push to your desired branch (main, staging, or dev).
   • Cloud Build will automatically use the cloudbuild.yaml file in the root of your repository.

2. Configure Substitutions and Secrets:

   • In the trigger settings, add required substitutions (e.g., _OPENAI_API_KEY, _DIFFBOT_API_KEY, etc.) as environment variables or use Secret Manager for sensitive data.

3. Push your code:

   • When you push to the configured branch, Cloud Build will build and deploy your backend (and optionally frontend) to Cloud Run using the steps defined in cloudbuild.yaml.

Manual Deployment

1. Set up Google Cloud SDK and authenticate:

   gcloud auth login
   gcloud config set project <YOUR_PROJECT_ID>
   

2. Run Cloud Build manually:

   gcloud builds submit --config cloudbuild.yaml \
     --substitutions=_REGION=us-central1,_REPO=cloud-run-repo,_OPENAI_API_KEY=<your-openai-key>,_DIFFBOT_API_KEY=<your-diffbot-key>,_BUCKET_UPLOAD_FILE=<your-bucket>,_BUCKET_FAILED_FILE=<your-bucket>,_PROJECT_ID=<your-project-id>,_GCS_FILE_CACHE=False,_TRACK_USER_USAGE=False,_TOKEN_TRACKER_DB_URI=...,_TOKEN_TRACKER_DB_USERNAME=...,_TOKEN_TRACKER_DB_PASSWORD=...,_TOKEN_TRACKER_DB_DATABASE=...,_DEFAULT_DIFFBOT_CHAT_MODEL=...,_YOUTUBE_TRANSCRIPT_PROXY=...,_EMBEDDING_MODEL=...,
       _EMBEDDING_PROVIDER=...,_BEDROCK_EMBEDDING_MODEL_KEY=...,_LLM_MODEL_CONFIG_OPENAI_GPT_5_2=...,_LLM_MODEL_CONFIG_OPENAI_GPT_5_MINI=...,_LLM_MODEL_CONFIG_GEMINI_2_5_FLASH=...,_LLM_MODEL_CONFIG_GEMINI_2_5_PRO=...,_LLM_MODEL_CONFIG_DIFFBOT=...,_LLM_MODEL_CONFIG_GROQ_LLAMA3_1_8B=...,_LLM_MODEL_CONFIG_ANTHROPIC_CLAUDE_4_5_SONNET=...,_LLM_MODEL_CONFIG_ANTHROPIC_CLAUDE_4_5_HAIKU=...,_LLM_MODEL_CONFIG_LLAMA4_MAVERICK=...,_LLM_MODEL_CONFIG_FIREWORKS_QWEN3_6=...,_LLM_MODEL_CONFIG_FIREWORKS_GPT_OSS=...,_LLM_MODEL_CONFIG_FIREWORKS_DEEPSEEK_V3=...,_LLM_MODEL_CONFIG_BEDROCK_NOVA_MICRO_V1=...,_LLM_MODEL_CONFIG_BEDROCK_NOVA_LITE_V1=...,_LLM_MODEL_CONFIG_BEDROCK_NOVA_PRO_V1=...,_LLM_MODEL_CONFIG_OLLAMA_LLAMA3=...
   

   • Replace the values in angle brackets with your actual configuration and secrets.
   • LLM_MODEL_CONFIG_FIREWORKS_QWEN3_6 is the app-facing config key for the fireworks_qwen3_6 model option and should map to the Fireworks serverless slug accounts/fireworks/models/qwen3p6-plus.
   • You can omit or add substitutions as needed for your deployment.

3. Monitor the build:

   • The build and deployment process will be visible in the Cloud Build console.

4. Access your deployed service:

   • After deployment, your backend will be available at the Cloud Run service URL shown in the Cloud Console.

---

Note:

• The cloudbuild.yaml file supports multiple environments (main, staging, dev) based on the branch name.
• The frontend build and deployment steps are commented out by default. Uncomment them in cloudbuild.yaml if you wish to deploy the frontend as well.

For more details, see the comments in cloudbuild.yaml.

---

Links

LLM Knowledge Graph Builder Application

Neo4j Workspace

Reference

Demo of application

Contact

For any inquiries or support, feel free to raise GitHub Issues

Happy Graph Building!