# Zoom MCP starter

This starter gives an AI client live Zoom tools plus a local SQLite ingestion lane.

## Setup

1. Copy `.env.example` to `.env` and add credentials from a Zoom Server-to-Server OAuth app.
2. Create a virtual environment: `python3 -m venv .venv`.
3. Install dependencies: `.venv/bin/pip install -r requirements.txt`.
4. Start the MCP server: `.venv/bin/python zoom_mcp.py`.
5. Register that absolute Python command and script path with your MCP client.

Never commit `.env` or the generated SQLite database.

## Ingest a month

```bash
.venv/bin/python ingest_zoom.py \
  --from-time 2026-07-01T00:00:00Z \
  --to-time 2026-08-01T00:00:00Z
```

The resulting `zoom-memory.sqlite` contains normalized meetings, summaries,
transcripts, sync state, and an FTS5 search index.