AI Book Recommender

A Domo Custom App that recommends books using Domo's AI Service Layer. The user picks favorite books (searched live against Open Library), genre, mood, and length; a structured system prompt drives the /domo/ai/v1/text/generation endpoint to return four JSON-formatted recommendations.

This repo is the companion sample for the AI Book Recommender tutorial.

What it demonstrates

• Scaffolding a Vite + React + TypeScript app with the DA CLI
• Debounced public-API search against Open Library
• Calling Domo's AI text generation endpoint with promptTemplate + system + outputWordLength
• Forcing structured JSON output with a strict system prompt (and a regex strip for fenced output)
• Building a polished Ant Design form and results view

Project layout

src/
├── main.tsx                       # App entry
├── index.scss
└── components/
    └── App/
        ├── App.tsx                # All state, prompts, API calls, and UI
        └── App.module.scss
public/
└── static/
    ├── bookshelf.jpeg             # Full-bleed background
    └── chapter_divider.png        # Heading divider

Prerequisites

• Node 18+
• The DA CLI and the Domo CLI
• domo login completed against your target instance
• Access to the Domo AI Service Layer on your instance

Getting started

1. Publish a skeleton to get a proxyId (required for authenticated Domo API calls):

   pnpm install
   pnpm upload

2. Create a card on the published design (no dataset needed — the card just exists to generate a proxyId) and copy the id + proxyId back into public/manifest.json.

3. Run locally:

   pnpm start

See the tutorial for the full walk-through, including the prompt design.

Scripts

Command	Description
pnpm start	Vite dev server with the Domo proxy
pnpm build	Lint, test, and build for production
pnpm upload	Build and domo publish in one step
pnpm generate	Scaffold new components / reducers with da generate
pnpm test	Run Vitest
pnpm storybook	Launch Storybook