Can I vibe code Crush.pics?

Build a lightweight Node.js and Express backend service designed to replace Crush.pics for Shopify stores. 1. DATA MODEL & STORAGE: Store job records in SQLite using Prisma. Schema: ImageJob (id, shopifyMediaId, originalUrl, compressedUrl, status, altText, originalSize, compressedSize, createdAt). Use AWS S3 or Cloudflare R2 bucket binding via AWS SDK for storing original uncompressed asset backups. 2. CORE WORKFLOW: Create a REST endpoint `/webhooks/shopify/media-updated` to receive Shopify `PRODUCTS_UPDATE` and `PRODUCTS_CREATE` webhooks. For each image URL found in product media: a) Download the high-res original image. b) Upload the original image to your S3/R2 backup bucket. c) Pass the buffer into `sharp` to convert to WebP/JPEG format with lossy quality=80 and progressive rendering. d) Send the optimized buffer back to Shopify using the GraphQL Admin API `stagedUploadsCreate` and update product media `fileUpdate` mutation. e) Generate ALT text based on the product title and variant options using a template string (e.g. '{product_title} - {variant_title}'), and update `alt` tag via GraphQL `productUpdate` mutation. 3. BULK QUEUE: Implement BullMQ with Redis to handle initial catalog optimization. Create an endpoint `/api/optimize-catalog` that fetches all product media via paginated GraphQL queries and queues processing jobs with concurrency limits to prevent hitting Shopify API rate limits. 4. FAILURE HANDLING: Prevent recursive execution loops by skipping images that already match compressed asset checksums. Include retry logic with exponential backoff for network drops. If image compression fails or increases file size, abort the mutation and retain the original image. OUT OF SCOPE: Embedded Shopify React dashboard UI, public multi-tenant billing, frontend app extensions.

$ each button prefixes agent-specific run instructions · build your own product, never copy proprietary code, trademarks or designs