Local Onboarding

Prerequisites

• Node.js 20+ (node --version)
• npm 10+ (npm --version)
• Ollama installed (for LLM + embeddings)
• 500 MB free disk space

Step 1: Install

git clone https://github.com/yourusername/docket.git
cd docket
npm install

Step 2: Run Setup

npm run setup

This interactive script will:

1. Create data/ directories
2. Copy config/config.example.yaml to config/config.yaml
3. Ask for your preferred LLM provider (Ollama, LM Studio, etc.)
4. Write the configuration file

Step 3: Verify Prerequisites

npm run doctor

Checks:

• Node.js version
• node_modules present
• Data directory writable
• Config file exists
• Ollama reachable

Step 4: Start Ollama

ollama pull llama3.2
ollama pull nomic-embed-text
ollama serve

Step 5: Start Docket

npm start

Visit http://localhost:3000/health to verify.

Hybrid Setup: LM Studio + Ollama

If you prefer LM Studio for chat and Ollama for embeddings, use this config:

docket:
  adapters:
    llm:
      default: "lm-studio"
      providers:
        lm-studio:
          adapter: "@docket/llm-openai-compatible"
          config:
            baseUrl: "http://localhost:1234/v1"
            model: "local-model"
    embedder:
      default: "ollama"
      providers:
        ollama:
          adapter: "@docket/embedder-ollama"
          config:
            baseUrl: "http://localhost:11434"
            model: "nomic-embed-text"
            dimensions: 768