GPT-4o SaaS Assistant Platform

I want to launch a desktop-first web app that lets users chat with an AI assistant through text, microphone input (with device selection), or image uploads, and watch answers appear in real time via GPT-4o streaming.

Usage is entirely time-based. Every newly-registered user receives a single 15-minute free session. Paid users can buy a 90-minute Pro session through Stripe Checkout; once a session starts, the server-side timer runs continuously until the minutes are gone—no pauses, no resets.

Accounts are created with traditional e-mail/password or an SSO flow (OpenID Connect or similar). An admin dashboard must give me at-a-glance insight into users, remaining minutes, active and past sessions, and completed Stripe payments, plus basic controls to deactivate or reset accounts.

Deliverables
• Desktop-optimised front end offering text, voice and image input, with GPT-4o streaming responses
• Secure back end that handles authentication, session timing, OpenAI calls and Stripe webhooks
• Admin dashboard with user, session and payment management
• Dockerfile / docker-compose and a concise README for local and production deployment
• Short hand-off walkthrough of the codebase and deployment steps

I will supply OpenAI and Stripe keys; you provide clean, well-commented code and any stack recommendations that best meet the requirements. When you reply, please highlight prior work with GPT streaming, Stripe integrations or similar metered SaaS models.