Raggadon CLI

NPM package for Raggadon RAG system - Provides project-based memory for AI assistants like Claude Code.

Installation

Global Installation

npm install -g raggadon
raggadon status

NPX (No Installation)

npx raggadon status
npx raggadon save "important information"

Quick Start

1. Initialize in your project:

   cd your-project
   npx raggadon init

2. Save important information:

   npx raggadon save "User authentication uses JWT tokens"
   npx raggadon save "API endpoint: POST /api/auth/login"

3. Search your project memory:

   npx raggadon search "authentication"
   npx raggadon search "API endpoints"

4. Check status:

   npx raggadon status

Commands

npx raggadon save <content>

Stores important information for the current project with semantic search capabilities.

Example:

npx raggadon save "The user model has fields: email, name, role, created_at"

npx raggadon search <query>

Searches for similar content in the current project using vector similarity.

Example:

npx raggadon search "user model"
# Returns relevant information about user-related code

npx raggadon status

Shows server status, project statistics, token usage, and costs.

Output includes:

• Server health status
• Current project name
• Operating mode (active/silent/ask)
• Memory entries count
• Monthly token usage and costs
• Recent activity

npx raggadon init

Initializes Raggadon for the current project by copying CLAUDE.md configuration.

npx raggadon mode <mode>

Changes the auto-save behavior mode.

Modes:

• active - Claude shows all RAG operations (default)
• silent - Works in background, check with status
• ask - Ask before each operation
• show - Display current mode

npx raggadon start

Starts the Raggadon server (if installed locally).

Requirements

• Raggadon Server: Must be running on http://127.0.0.1:8001
• Node.js: Version 16 or higher
• Backend: Supabase with pgvector + OpenAI API

How It Works

1. Project Detection: Uses current directory name as project identifier
2. Content Storage: Converts text to embeddings using OpenAI's text-embedding-3-small
3. Vector Search: Uses PostgreSQL + pgvector for similarity search
4. Cost Tracking: Monitors token usage and estimates costs

Cost Structure

• Model: text-embedding-3-small (1536 dimensions)
• Rate: $0.02 per 1,000 tokens
• Typical save: ~150-200 tokens (~$0.000003-0.000004)
• Monthly budget: Usually under $0.01 for normal usage

Claude Code Integration

When you run npx raggadon init, it adds a CLAUDE.md file that instructs Claude Code to:

• Automatically save important code definitions
• Search project memory before answering questions
• Store architecture decisions and configurations
• Track TODOs and bug fixes

Architecture

┌─────────────────┐    HTTP API    ┌──────────────────┐
│  NPX Package    │ ──────────────→ │  Raggadon Server │
│  (CLI Client)   │                 │  + Supabase      │
│                 │ ←────────────── │  + OpenAI        │
└─────────────────┘                 └──────────────────┘

Examples

Development Workflow:

# Start new project
cd my-awesome-app
npx raggadon init

# Save key information as you code
npx raggadon save "Uses React 18 with TypeScript and Vite"
npx raggadon save "API endpoints defined in src/api/routes.ts"
npx raggadon save "Authentication middleware in src/middleware/auth.ts"

# Search when you need context
npx raggadon search "authentication"
npx raggadon search "API structure"

# Check your usage
npx raggadon status

Team Collaboration:

# Each team member can access shared project memory
npx raggadon search "deployment process"
npx raggadon search "database schema"

Troubleshooting

Server not running:

npx raggadon status
# Shows: ❌ Raggadon läuft nicht!

Solution: Start the Raggadon server or check your installation.

No results found:

• Check if you're in the correct project directory
• Verify you've saved information with npx raggadon save
• Try broader search terms

License

MIT

Repository

https://github.com/haltever/raggadon