The 5i Consortium

5 intelligences. One synthesized answer.

A multi-model AI writing synthesis engine by Creative Konsoles. Paste a message, set your intent and tone, and run it through up to 5 AI models simultaneously. Every model contributes a distinct perspective. A synthesis layer combines the best of all of them into one superior output.

Nothing like this exists at this price point.

---

The Problem

Every major AI model has a different voice, a different blindspot, and a different strength. Claude is nuanced. GPT-4 is decisive. Gemini is comprehensive. Grok is unfiltered. Mistral is surgical. Running one model means you get one perspective — and you don't know what you're missing.

The 5i Consortium runs all five in parallel and synthesizes the output. You get the best version of what you were trying to say, not just the first one.

---

Features

• ⚡ Parallel model execution — all active models run simultaneously, not sequentially
• 🧠 Synthesis layer — Claude distills every model's output into one final answer
• 🎛 Intent presets — Rewrite, Summarize, Improve, Tone Check, Draft Reply, Analyze, Proofread, Soften, Strengthen, Less AI
• 🎚 Tone fader — casual to formal, continuously variable
• 👤 Context tags — Boss, Client, Coworker, Friend, General
• 📋 Model breakdown — collapsible view of every model's raw contribution
• 🔒 Locked model previews — blurred teasers of what you're missing on free tier
• 🕑 Local history — last 20 syntheses saved to localStorage, tap to restore
• 📱 PWA-ready — installable on iOS and Android, mobile-first layout
• 🌑 Dark theme — logo fills the entire screen, UI floats on glass

---

How It Works

User input (message + intent + tone + context)
  │
  ├── /api/run → Claude          (Anthropic SDK, claude-sonnet-4-6)
  ├── /api/run → GPT-4           (OpenAI SDK, gpt-4o)
  ├── /api/run → Gemini          (Google Generative AI SDK, gemini-2.0-flash)
  ├── /api/run → Mistral         (OpenAI-compatible, mistral-large-latest)
  └── /api/run → Grok            (OpenAI-compatible, x.ai endpoint)
            │
            │  All run in parallel via Promise.all
            ▼
  /api/synthesize → Claude       (combines all outputs into one)
            │
            ▼
  SynthesisCard                  (primary output — the answer)
  ModelBreakdown                 (collapsible, shows each model's contribution)

Each model runs with a distinct persona prompt — not just the same system message sent five times. Claude reasons for depth. GPT-4 cuts for efficiency. Gemini reads subtext. Grok strips the polish. Mistral operates with surgical precision.

---

Tech Stack

• Next.js 16 (App Router) + TypeScript
• Tailwind CSS
• Anthropic SDK — Claude Sonnet 4.6
• @google/generative-ai — Gemini 2.0 Flash
• OpenAI SDK — GPT-4o, Grok (x.ai), Mistral (OpenAI-compatible)
• Vercel — deployed, serverless API routes

---

Model Status

Model	Provider	Status
Claude	Anthropic	✅ Live
GPT-4	OpenAI	✅ Live
Gemini	Google	✅ Live
Mistral	Mistral AI	✅ Live
Grok	xAI	⏳ Pending credits

---

Quick Start

git clone https://github.com/papjamzzz/The-5i-Consortium.git
cd The-5i-Consortium
npm install

Create .env.local with your API keys:

ANTHROPIC_API_KEY=your-key
OPENAI_API_KEY=your-key
GEMINI_API_KEY=your-key
MISTRAL_API_KEY=your-key
GROK_API_KEY=your-key

npm run dev

Open http://localhost:3000

---

Pricing Tiers

Tier	Price	Access
Free	$0	3 syntheses/day · Claude only
Base	$19/mo	Unlimited · 3 models · full Intelligence Report
Premium	$49/mo	All 5 models · full signal stack · chain · blend · AI detection · humanize · API
Founding Member	$149 once	Lifetime Premium + founding badge

---

Project Structure

The-5i-Consortium/
├── app/
│   ├── page.tsx                  ← Full UI (Compose, Library, History, Account)
│   ├── layout.tsx                ← Root layout + metadata
│   ├── globals.css               ← Design system + glass surfaces
│   └── api/
│       ├── run/route.ts          ← Per-model API route (all 5 models)
│       └── synthesize/route.ts   ← Synthesis layer (Claude)
├── public/
│   ├── logo.png                  ← Full-screen background
│   ├── manifest.json             ← PWA config
│   └── icons/                   ← iOS + Android app icons
└── .env.local                    ← API keys (not committed)

---

Roadmap

• Clerk auth (email + Google OAuth)
• Stripe subscriptions (Essential, Pro, Founding Member)
• Supabase usage tracking
• Upstash Redis rate limiting
• Chain Mode — sequential model refinement
• Blend Mode — weighted multi-model fusion
• AI Detection Score (GPTZero integration)
• Humanize pass

---

License

MIT

---

Built by Creative Konsoles. The Consortium runs so you don't have to choose.