01
Image
Flux, SDXL and DALL·E 3 — with automatic fallback between providers.
Personal Project — AI Platform / SaaS
Genora
One prompt. Infinite possibilities.
Genora is an all-in-one AI creation platform: a single prompt becomes an image, video, avatar, voice or song, all inside one account with real credits, plans and billing — designed, architected and built solo.
AI PlatformFull-stackMulti-model
What Genora generates
02
Video
RunwayML Gen-3 to turn a prompt and image into a clip.
03
Avatar
Avatars generated from a single reference photo.
04
Voice
Text-to-speech with ElevenLabs and OpenAI TTS.
05
Music
Tracks generated from a style and mood description.
Architecture & technical decisions
The hardest part wasn't generating content with AI — it was doing it reliably when any external provider can fail at any moment.
Automatic provider fallback
If Replicate fails on an image request, it automatically falls back to fal.ai — the user never sees the error.
Prompt enhancement with GPT-4o-mini
Every prompt is automatically refined before reaching the image or video model, for richer results.
Real-time status via SSE
Async generations (video, avatar) stream their status via Server-Sent Events, no manual polling on the front end.
Credits & billing with Stripe
Every model has a credit cost; subscriptions, webhooks and balance are synced through Stripe.
Tech stack
Next.js 15 (App Router)Supabase (Postgres + Auth + Storage)StripeZustandReact QueryReplicate · fal.ai · OpenAI · ElevenLabs · RunwayML
Business model
Free
Free
50 credits/mo
Pro
R$ 59/mo
500 credits/mo
Business
R$ 199/mo
2,000 credits/mo
Deliverables
- —Full-stack architecture
- —Multi-AI orchestration
- —Credits system
- —Stripe integration
- —Auth & middleware
- —Real-time status (SSE)
My role
Product architecture, backend, integration with multiple AI providers and front-end — solo, from scratch.