Foundation

The Foundation module sets up the base configuration for all other modules. You'll configure your Obsidian vault structure and connect your preferred AI provider.

Prerequisites

  • Obsidian installed and a vault created
  • Python 3.11+ installed
  • ~15 minutes to complete

What You'll Get

  • Organized folder structure for meetings, transcripts, and templates
  • Central configuration file for all scripts
  • AI provider connected and tested
  • Ready to add additional modules

Vault Structure

Create these folders in your Obsidian vault:

YourVault/
├── Meetings/           # Processed meeting notes
├── Transcripts/        # Archived raw transcripts
├── People/             # Person notes (@ First Last.md)
├── Templates/          # Tag reference and templates
└── scripts/            # Processing scripts

AI Provider Setup

Choose one of the following AI providers. Each has different trade-offs:

Ollama (Local - Maximum Privacy)

Best for: Users who want all data to stay on their machine.

  1. Download Ollama from ollama.ai
  2. Install and start Ollama
  3. Pull a model: ollama pull llama3.2
  4. Verify it's running: curl http://localhost:11434/api/tags
# config.yaml
ai_provider: ollama
ai_model: llama3.2
ai_endpoint: http://localhost:11434

OpenAI API (Easiest Setup)

Best for: Users who want quick setup and reliable performance.

  1. Get an API key from OpenAI Platform
  2. Set your API key as an environment variable
# Set in your shell profile (.zshrc, .bashrc)
export OPENAI_API_KEY="sk-..."
# config.yaml
ai_provider: openai
ai_model: gpt-4o-mini
ai_endpoint: https://api.openai.com/v1

Anthropic Claude API

Best for: Users who prefer Claude's writing style.

  1. Get an API key from Anthropic Console
  2. Set your API key as an environment variable
# Set in your shell profile
export ANTHROPIC_API_KEY="sk-ant-..."
# config.yaml
ai_provider: anthropic
ai_model: claude-3-5-sonnet-20241022
ai_endpoint: https://api.anthropic.com

Google Vertex AI (Enterprise)

Best for: Enterprise users with Google Cloud access.

  1. Set up a Google Cloud project with Vertex AI enabled
  2. Configure authentication with gcloud CLI
# config.yaml
ai_provider: vertex
ai_model: gemini-1.5-flash
ai_endpoint: us-central1-aiplatform.googleapis.com
gcp_project: your-project-id

Configuration File

Create config.yaml in your scripts folder:

# config.yaml - AI Executive Assistant Configuration

# Vault paths (relative to vault root)
vault_path: ~/Documents/MyVault
meetings_folder: Meetings
transcripts_folder: Transcripts
people_folder: People
templates_folder: Templates

# AI Provider (ollama, openai, anthropic, vertex)
ai_provider: ollama
ai_model: llama3.2
ai_endpoint: http://localhost:11434

# Processing preferences
summary_style: concise  # concise or detailed
action_item_format: "- [ ] @Owner — Task"
date_format: "%Y-%m-%d"

Verify It Works

Test your AI provider connection:

python scripts/test_connection.py

Expected output:

✓ Config loaded successfully
✓ AI provider connected (ollama)
✓ Test prompt completed in 1.2s
Ready to process meetings!

Troubleshooting

Ollama not responding

Make sure Ollama is running: ollama serve

API key not found

Verify your environment variable is set: echo $OPENAI_API_KEY

Python module not found

Install dependencies: pip install -r requirements.txt

Next Steps

Now that your foundation is set up, continue to Meeting Processing to enable AI-powered summaries.