TikTok Shop Fulfillment Automation

AI Prompts to Build This

Copy these build prompts into Claude, Cursor, or your AI coding tool. Create a free account to unlock the full research behind them.

1. 1. Project Setup

   Build the weekend MVP of "ShopAutopilot": a TikTok Shop seller imports orders and stock from CSV exports, and the app flags oversold items, buys shipping labels in one batch and exports tracking numbers to paste back. Labels use a test key until you trust it.
   
   Stack: Next.js (App Router), TypeScript, Tailwind, Supabase (Postgres, Row Level Security, Auth with email magic link), the EasyPost API for rates and labels (test mode first). Orders and inventory come in as the CSV files that Seller Center exports. Deploy on Vercel.
   
   Tables (Row Level Security on, each user reads only their own rows):
   - imports(id, user_id, kind, filename, row_count, created_at) kind is orders or inventory
   - products(id, user_id, sku, name, stock_qty)
   - orders(id, user_id, tiktok_order_id, status, buyer_name, address jsonb, items jsonb, region, carrier, tracking_number, label_url, created_at, shipped_at) tiktok_order_id is unique per user
   - sync_errors(id, user_id, order_id, error jsonb, created_at)
   
   Screens: /login, /import, /orders (select and ship), /inventory (stock and low-stock badges), /export (the tracking CSV), /calculator (public).
   Env vars (names only): NEXT_PUBLIC_SUPABASE_URL, NEXT_PUBLIC_SUPABASE_ANON_KEY, SUPABASE_SERVICE_ROLE_KEY (server only), EASYPOST_API_KEY.
   
   Do not build: billing, plans or order-count tiers, the TikTok Shop API or OAuth, a queue or Redis, carrier aggregator accounts, an analytics tab.
   Done when: npm run dev starts, you can sign in, and the four tables exist with Row Level Security on.

2. 2. Core Feature

   Build the one feature that proves ShopAutopilot: a spike of orders turns into labels and tracking numbers in a few clicks, without selling stock you do not have.
   1. Import an orders CSV and an inventory CSV from Seller Center. Insert orders idempotently by tiktok_order_id so a re-import changes nothing. Show rows that failed to parse and why.
   2. When an order is imported, reduce stock for each SKU. Flag any order whose quantity is more than the stock on hand as oversold and keep it out of the batch.
   3. The orders page lists pending orders with a select-all and a Ship button. Group the selected orders by destination region and the seller's carrier preference.
   4. For each group, create shipments and buy labels through the EasyPost API, a few at a time, retrying a rate limit with backoff. Save the tracking number, carrier and label URL on each order, set shipped_at and show a printable label sheet.
   5. Record every failure in sync_errors with the raw error and leave that order pending. A Retry button reruns only the failed ones.
   6. The export page downloads a CSV of order id, carrier and tracking number to paste back into Seller Center.
   7. Delete a buyer's address 30 days after shipping, checked whenever the orders page loads.
   8. The public calculator takes daily orders and minutes per order by hand and shows hours saved a week and the value at an hourly rate the visitor enters, with the formula shown.
   Rules: use the EasyPost test key until a live label is bought on purpose. Never buy a label for an oversold order. Never log a full address.
   Empty state: with no orders, show the two CSV formats and sample files.
   Done when: importing a sample of 50 orders twice creates 50 orders, an oversold order is flagged and excluded, a batch of 20 gets labels with tracking numbers, a forced EasyPost error lands in sync_errors and the export lists exactly the shipped orders.

3. 3. Landing Page

   Build a one-page landing site for ShopAutopilot, fulfillment help for TikTok Shop sellers.
   Hero: "Ship the spike without the spreadsheet." Sub: "Import your orders, catch oversold items, buy labels in one batch and get tracking numbers ready to paste back." One button: Join the waitlist.
   Sections: an efficiency calculator (daily orders and minutes per order in, hours saved out, with the formula shown), how it works in four steps (import, check stock, batch ship, export tracking), and an FAQ on which exports are supported, what is coming next (a direct connection to TikTok Shop) and how buyer addresses are handled (deleted 30 days after shipping).
   Waitlist: store the email in a waitlist table in Supabase. No other service.
   Style: Geist, a clean high-contrast look with dark navy and one orange accent, no illustrations of people.
   Done when: the page renders on a phone and a submitted email appears in the waitlist table.

4. 4. Branding Package

   Use a design or image tool for this one. A coding agent cannot draw a logo.
   Brand for ShopAutopilot: a wordmark and an icon that suggest a parcel with an arrow that has already left.
   Colors: dark navy, off-white and one orange. Type: Geist, with tabular numbers for counts.
   Deliverables: wordmark, icon, four order status chips (pending, oversold, shipped, error) that differ by shape as well as color, a label sheet header, and one launch graphic.
   Done when: each deliverable is saved in one folder and the status chips are distinguishable without color.