ViaProto

Your AI learning architect — describe a goal, get a structured path from fundamentals to mastery, each step backed by a real, verified resource.

🌐 Live at https://viapro.to

---

What It Does

ViaProto solves the problem of information overload and deteriorating search quality. Instead of spending hours wading through SEO-gamed content farms, users get:

• Structured learning paths with 5-8 sections of curated resources
• 40+ AI models to choose from (Claude, GPT, Gemini, DeepSeek, etc.)
• Verified resources via AI web search (videos, articles, books, projects, courses, etc.)
• Free & paid options clearly labeled for each resource
• Team collaboration with seat-based pricing

Target audience: Professionals who need to upskill quickly in an era of mounting layoffs and increasing automation.

---

Tech Stack

Category	Technology	Version
Framework	Next.js (App Router)	15.1.8
Language	TypeScript	5.9.2
UI	React	19.0.0
Styling	Tailwind CSS + DaisyUI	4.1.10 + 5.0.5
Database	Supabase (PostgreSQL)	2.45.0
CMS	TinaCMS (blog)	3.9.3
Auth	Supabase Auth	2.45.0
AI	OpenRouter (40+ models)	via OpenAI SDK 6.9.1
Payments	Stripe	13.11.0
Email	Resend	4.0.1
Images	Unsplash API	Custom integration
Background Jobs	Graphile Worker	0.16.6
Deployment	Coolify → Hetzner VPS	-

---

Quick Start

Prerequisites

• Node.js 18+
• Supabase CLI (install)
• Accounts: Supabase, Stripe, OpenRouter, Resend, Unsplash

Installation

# Clone and install
git clone https://github.com/chznbaum/via-proto.git
cd via-proto
npm install

# Environment setup
cp .env.example .env.local
# Edit .env.local with your keys

Database Setup

# Start local Supabase
supabase start

# Apply migrations and seed data
supabase db reset

# Generate/apply topic seeds
node scripts/generate-seeds.js

Update .env.local with the credentials from supabase start output.

Run Development Server

# Terminal 1: Next.js dev server
npm run dev
# → http://localhost:3001

# Terminal 2: Background job worker (required for path generation)
npm run worker:dev

---

Environment Variables

# Supabase
NEXT_PUBLIC_SUPABASE_URL=
NEXT_PUBLIC_SUPABASE_ANON_KEY=
SUPABASE_SERVICE_ROLE_KEY=
DATABASE_URL=                  # Postgres connection string (required by the worker)

# Stripe
STRIPE_PUBLIC_KEY=
STRIPE_SECRET_KEY=
STRIPE_WEBHOOK_SECRET=

# AI
OPENROUTER_API_KEY=

# Services
RESEND_API_KEY=
UNSPLASH_ACCESS_KEY=

# Site
NEXT_PUBLIC_SITE_URL=http://localhost:3001

# Worker (Background Jobs)
WORKER_CONCURRENCY=3
WORKER_POLL_INTERVAL=1000

# TinaCMS (blog)
NEXT_PUBLIC_TINA_CLIENT_ID=    # Client ID from TinaCloud
TINA_TOKEN=                    # Read-only content token
TINA_SEARCH_TOKEN=             # Search indexer token

---

Project Structure

via-proto/
├── app/
│   ├── (auth)/               # Auth route group — login, register, OAuth callback
│   ├── (dashboard)/          # Protected route group — dashboard, skills, progress, account, upgrade
│   ├── (main)/               # Public route group (own root layout)
│   │   ├── blog/             # Blog (TinaCMS-managed MDX)
│   │   ├── explore/          # Public path discovery
│   │   ├── paths/[id]/       # Public path detail page
│   │   ├── invite/           # Team invite acceptance
│   │   └── layout.tsx        # App layout (html/body, fonts, analytics)
│   ├── api/                  # API route handlers
│   │   ├── paths/            # Path generation & management
│   │   ├── stripe/           # Checkout & portal
│   │   ├── topics/           # Topic search
│   │   └── webhook/          # Stripe webhooks
│   └── layout.tsx            # Root layout (metadataBase)
├── components/               # React components
├── contexts/                 # React context providers
├── hooks/                    # Shared React hooks
├── content/
│   └── posts/                # Blog posts as MDX (managed by TinaCMS)
├── tina/
│   └── config.ts             # TinaCMS schema & configuration
├── libs/
│   ├── models/               # AI model configuration (40+ models)
│   ├── jobs/                 # Background job tasks (Graphile Worker)
│   ├── supabase/             # Database clients (server, client, service)
│   ├── openrouter.ts         # AI generation logic
│   ├── stripe.ts             # Payment utilities
│   └── validation/           # Zod schemas
├── supabase/
│   ├── migrations/           # Database schema (timestamped)
│   └── seed.sql              # Topic seed data
├── data/                     # JSON source for seeds
├── scripts/
│   └── generate-seeds.js     # Seed file generator
├── worker.ts                 # Background worker entry point
└── config.ts                 # App configuration

Note on Route Groups: The app uses three route groups, each with its own layout:

• (main) - Public-facing pages with custom fonts, analytics, and client-side providers
• (dashboard) - Authenticated/protected pages (dashboard, skills, progress, account, upgrade)
• (auth) - Login, register, and OAuth callback flows

---

Key Concepts

Rate Limits by Tier

Tier	Monthly Paths	Cost	Models
Free	1	$0	7 free models
Pro	5	$12/mo or $100/yr	All 40+ models
Team	10 + (3 per seat)	$10/seat/mo	All 40+ models

Team scaling: 2 seats = 10 paths, 5 seats = 19 paths, 10 seats = 34 paths

Path Generation Flow (Background Jobs)

1. User selects topic, skill level, optional goals, optional AI model
2. System checks rate limits and model access
3. Job #1 (generate_metadata): AI generates title, description, skill level
4. Job #2 (fetch_unsplash_image): Fetches cover image from Unsplash
5. Job #3 (generate_sections_resources): AI generates 5-8 sections with 3-7 resources each
6. Multi-stage status tracking: pending → generating_metadata → fetching_image → curating_resources → completed
7. Each job retries up to 3 times on failure; users notified via email on permanent failure
8. Result: 20-100+ hours of curated learning content

Worker Process: Run npm run worker to start the Graphile Worker background processor

Account Architecture

• Personal accounts: 1 user, 1 seat
• Team accounts: 2+ users, shared quota
• Learning paths belong to accounts, not users
• Billing via Stripe customer_id on account

Database Schema

Core tables with Row-Level Security:

• profiles - User profiles
• accounts - Billing entities (personal/team)
• account_users - User ↔ Account relationship
• topics - Pre-seeded learning topics
• learning_paths - AI-generated paths
• sections - Path sections (ordered)
• resources - Learning resources (videos, articles, etc.)
• unsplash_images - Cached featured images

---

Development

Common Commands

npm run dev          # Development server (port 3001)
npm run worker       # Background job worker (required for path generation)
npm run worker:dev   # Worker with hot reload
npm run build        # Production build
npm run lint         # ESLint check
npm run postbuild    # Generate sitemap (auto-runs)

# Database
supabase db reset    # Reset with migrations + seed
supabase db push     # Push migrations to remote

# Stripe testing
stripe listen --forward-to localhost:3001/api/webhook/stripe

Add a New AI Model

Edit libs/models/model-config.ts and add to MODEL_CATALOG:

{
  id: "provider/model-name",
  name: "Display Name",
  provider: "ProviderName",
  minimumTier: "free" | "pro",
  costTier: "free" | "low" | "medium" | "high" | "premium",
  supportsWebSearch: true,
  supportsStructuredOutput: true,
  description: "Brief description",
  featured: true,  // Optional
}

Model appears automatically in UI and generation API.

---

Deployment

Deployed via Coolify to Hetzner VPS with BunnyCDN for DNS/CDN.

Coolify Setup

Two Services Required:

Service 1: Web (Next.js)

1. Create Next.js application in Coolify
2. Set build command: npm run build
3. Set start command: npm run start
4. Add production environment variables
5. Link Supabase remote database
6. Configure domain: viapro.to

Service 2: Worker (Background Jobs)

1. Create Node.js application in Coolify
2. Set build command: npm install
3. Set start command: npm run worker
4. Share same environment variables as web service
5. Set restart policy: always
6. Both services must connect to same DATABASE_URL

Production Checklist

• Apply database migrations: supabase db push
• Seed topics database
• Configure Stripe webhook: https://viapro.to/api/webhook/stripe
• Test payment flows
• Verify RLS policies active
• Test path generation
• Enable SSL (Coolify auto-handles)

Full deployment guide: https://coolify.io/docs/applications/nextjs

---

API Endpoints

Path Generation

• POST /api/paths/initiate - Create path and queue background job
• GET /api/paths/[id]/status - Check generation status (poll this during generation)
• GET /api/paths/[id] - Get path details

Other

• POST /api/stripe/create-checkout - Start payment
• POST /api/stripe/create-portal - Manage subscription
• POST /api/webhook/stripe - Handle Stripe events
• GET /api/models - List available AI models
• GET /api/topics - Search topics (typeahead)

---

Support

• Support: support@viapro.to
• Privacy: privacy@viapro.to
• Security: security@viapro.to

---

License

Proprietary - All rights reserved. Not open source.

---

Built with Next.js, Supabase, OpenRouter, Stripe, and Tailwind CSS.