Installation

Requirements

• Node.js 18+ (for the Node adapter and local development)
• npm 9+ (for workspaces support)

Individual runtimes have their own requirements:

• AWS Lambda — Node.js 18+ runtime
• Cloudflare Workers — compatibility check via wrangler
• Bun 1.x or Deno 2.x — no adapter needed, app.fetch matches natively

Quick install (scaffold)

The fastest way to start is the interactive scaffolding CLI:

npm create nodalite
# or
npx nodalite create

It generates a project with all necessary dependencies pre-configured. See the Scaffolding guide for details.

Manual install

Install the core library:

npm install nodalite
# or the scoped form:
npm install @nodalite/core

Then add adapters as needed:

# Adapters (choose your runtime)
npm install @nodalite/adapter-node
npm install @nodalite/adapter-lambda
npm install @nodalite/adapter-edge

# Background workers (Node-only)
npm install @nodalite/workers

# Scheduler (Node-only)
npm install @nodalite/scheduler

# ML inference (optional, needs onnxruntime-node or custom engine)
npm install @nodalite/ml

Monorepo setup (for contributors)

Clone the repo and install everything:

git clone https://github.com/AkkilMG/nodalite.git
cd nodalite
npm install

Build every package:

npm run build --workspaces --if-present

Run all tests:

npm test

Optional: ONNX native dependency

@nodalite/ml's onnxEngine() requires the onnxruntime-node package (~270MB native binary). It's a peer dependency — install it only if you need it:

npm install onnxruntime-node

If you don't, Model still works with any custom InferenceEngine implementation, or you can use the WASM backend instead. See ML Inference.