Customer Reviews to Personas to Image Ads

Copy this prompt & paste into Claude Code

Build an internal tool that turns customer reviews into persona-specific static ads using the Static Ads Lab API.

Turn a customer reviews export (4-5 star reviews) into buyer personas with Gemini, then batch-create audiences and Meta-ready image ads.

Stack: TypeScript, Express, Vercel AI SDK, Google Gemini, Static Ads Lab API.

View on GitHub

What it builds

A small Express server with a vanilla UI. Upload a CSV of reviews, infer buyer personas with AI, select personas, then generate one image ad per persona × template combination.

Key patterns

Batch image ad creation with parallel POST /v1/image-ads calls
Batched status polling via GET /v1/image-ads?ids=... (one request per tick, not N)
Server-only API keys — the browser never touches secrets
localStorage for generation session persistence across page refreshes

Prerequisites

Node.js 20+
A Google AI Studio key for Gemini
A Static Ads Lab API key
Brand, product, and design template IDs already set up in your workspace

Quick start

cd recipes/reviews-to-persona-image-ads
npm install
cp .env.example .env
# Edit .env and src/config/recipe-constants.ts with your IDs
npm start

Open http://localhost:9000 — paste or upload a CSV, infer personas, select the required number, and generate ads.

Each recipe includes a prompt.md file designed to be handed directly to Claude Code, Cursor, or your preferred AI coding assistant. It contains the project's intent, architecture, edit points, and suggested next tasks.