Skip to content

Getting Started

This guide covers prerequisites, installation, and verification for MegaBrain.

Prerequisites

Backend Prerequisites

  • Java 22 or higher 

    java -version  # Should show 22 or higher

  • Maven 3.8 or higher 

    mvn -version  # Should show 3.8 or higher

  • PostgreSQL 12+ (optional, for vector search) 

    psql --version

  • Neo4j 5.x (optional, for graph database) 

    neo4j version

  • Ollama (optional, for local LLM) 

    ollama --version
# Or install from https://ollama.ai
    For offline operation: run Ollama locally and pull models before disconnecting (e.g. ollama pull codellama). Inference uses only the configured endpoint; no internet is required at runtime.

Frontend Prerequisites

  • Node.js 18+ and npm 

    node --version  # Should show 18 or higher
npm --version

  • Angular CLI 20 (install globally) 

    npm install -g @angular/cli@20
ng version

Installation Steps

1. Clone the Repository

git clone <repository-url>
cd MegaBrain

2. Backend Setup

cd backend

# Verify Java and Maven
java -version
mvn -version

# Compile the project
mvn clean compile

# Run tests
mvn test

# Start in development mode
mvn quarkus:dev

The backend will start on http://localhost:8080

3. Frontend Setup

cd frontend

# Install dependencies
npm install

# Start development server (with API proxy)
npm start
# or
ng serve

The frontend will start on http://localhost:4200 and proxy API requests to http://localhost:8080

4. Verify Installation

Backend Health Check: 

curl http://localhost:8080/q/health

Expected response: 

{
  "status": "UP",
  "message": "MegaBrain is running"
}

Frontend: Open http://localhost:4200 in your browser. You should see the MegaBrain dashboard.

5. CLI (optional)

When the backend is built for CLI mode, you can run the MegaBrain CLI. The ingest and search commands are available; use megabrain ingest --help or megabrain search --help to see usage and options. The search command supports filter options (--language, --repo, --type, --limit) and output options (--json, --quiet, --no-color); see CLI Reference for details. Use --json for scripting (e.g. megabrain search "query" --json or --json --quiet for the results array only). When you run an ingest (e.g. megabrain ingest --source github --repo owner/repo), progress is streamed in the terminal. Use --verbose for detailed progress and stack traces on errors.

cd backend
mvn package
java -jar target/quarkus-app/quarkus-run.jar ingest --help

6. Local LLM (Ollama) – offline operation

To use the local LLM without internet connectivity (AC3):

  1. Install and start Ollama on the same machine or a reachable host.
  2. Pull the model you will use while online: ollama pull codellama (or mistral, llama2, etc.).
  3. Configure MegaBrain to use that endpoint (default is http://localhost:11434) and the same model name in application.properties (see Configuration Reference).
  4. At runtime, all LLM requests go only to the configured Ollama endpoint; no external API calls are made. You can disconnect the network and continue using the local LLM.

Next Steps