Add OpenAI (GPT) Features to an Existing React + Node.js Auto-Parts PDF App

We have a working full-stack web app that lets automobile parts exporters input car + parts data and generate clean, multi-page PDF inspection/quotation sheets. The app is built and functional — we now want to upgrade it with AI features powered by the OpenAI API. Current tech stack (already built): Frontend: React + TypeScript, Tailwind CSS, shadcn/ui Backend: Node.js + Express, MongoDB (Mongoose), JWT auth PDF: Puppeteer (HTML → PDF) Images: Cloudinary Clean layered architecture (routes → controllers → services → repositories), Zod validation What we need (AI integration): We want to add GPT-powered features. Final scope is flexible, but the priorities are: AI-generated part descriptions — generate professional, multilingual product descriptions from a part code / name to include in the PDF. Smart part-info autofill — suggest part name, category, and details from a 12-digit part code or uploaded image. (Nice to have) Image recognition — identify the part from its photo (OpenAI vision) and pre-fill fields. (Nice to have) Quotation summary / notes — auto-draft a professional summary or cover note for the document. You'll integrate the OpenAI API cleanly into the existing backend service layer (with proper error handling, rate limiting, and key management), and add the matching UI in React. Deliverables: OpenAI API integrated into the backend with the agreed features React UI for the new AI features (loading states, error handling, editable AI output) Secure key handling via environment variables Sensible cost controls (token limits, caching where it makes sense) Brief documentation of what was added and how to configure it Required skills: Strong React + TypeScript and Node.js + Express Hands-on experience with the OpenAI API (chat completions; vision a plus) Comfortable working in an existing codebase (not greenfield) MongoDB, REST APIs, clean code practices Project details: Timeline: [e.g. 1–2 weeks] Codebase shared with the selected freelancer (private repo) To apply, please answer: Have you integrated the OpenAI API into a production app before? Share an example. Which approach would you use for AI part descriptions, and how would you keep token costs low? Roughly how long would the priority features (1 & 2) take you? Please start your proposal with the word "PARTS" so I know you read this.