Skip to main content

Scaffolding Projects

vai scaffold creates a complete, runnable starter project with all files needed for a Voyage AI + MongoDB Atlas Vector Search integration.

Quick Start​

# Node.js project
vai scaffold my-search-app

# Next.js project with search UI
vai scaffold my-search-ui --target nextjs

# Python Flask API
vai scaffold my-api --target python

What Gets Created​

vanilla (Node.js)​

  • lib/voyage.js — Voyage AI client
  • lib/mongodb.js — MongoDB connection
  • lib/retrieval.js — Vector search + rerank
  • lib/ingest.js — Document ingestion
  • lib/search-api.js — Express search endpoint
  • package.json, README.md, .env.example

nextjs (Next.js)​

  • lib/voyage.js, lib/mongodb.js — Client libraries
  • app/api/search/route.js — Search API route
  • app/api/ingest/route.js — Ingest API route
  • app/search/page.jsx — Search page with UI
  • lib/theme.js, app/layout.jsx — MUI theme and layout

python (Python)​

  • voyage_client.py — Voyage AI client
  • mongo_client.py — MongoDB connection
  • app.py — Flask app with search routes
  • chunker.py — Document chunking
  • requirements.txt, README.md

Customizing​

vai scaffold my-app --target nextjs \
  --model voyage-4 \
  --db myapp \
  --collection knowledge \
  --dimensions 512

Further Reading​