Yonoo Engine

Internal documentation for founders and investors

Invalid access code

The Yonoo Engine

How we turn one question into the best possible answer โ€” using every frontier AI model available. The complete decision tree from user input to synthesized output.

Why Yonoo Exists

๐ŸŽฏ

One Subscription. Every AI.

ChatGPT + Claude + Perplexity + Gemini + Grok + DeepSeek + Llama + GLM = โ‚ฌ77+/mo separately. Yonoo Pro = โ‚ฌ19.99/mo.

๐Ÿง 

You Don't Pick. We Route.

Users don't need to know which AI is best. Yonoo's intelligent router analyzes every question and sends it to the optimal model automatically.

๐Ÿ”ฎ

Yonoo Fusion

When accuracy matters most: send to multiple AIs simultaneously, synthesize the best answer. Inspired by Perplexity's Model Council, included in every plan.

๐ŸŽ“

Built for Students & Families

Starting at โ‚ฌ5.99/mo. Study modes, quizzes, family plans for 5 users at โ‚ฌ29.99. No AI company serves this market specifically.

The Complete Flow: Question to Answer

STEP 1

User Sends a Question

From the chat interface at yonoo.ai/app. The message includes: text content, selected task type (Write, Learn, Create, etc.), active Composer Mode (Fusion, Coding, Research, etc.), conversation history, and any uploaded files.

STEP 2

Authentication & Rate Limiting

Server validates user token, checks subscription tier (Free/Student/One/Pro/Family), and enforces usage limits. Free: 50 chats/week. Student: 100/week. Pro: unlimited. All tiers get access to all 8 AI models โ€” we democratize access, not quality.

STEP 3

Special Mode Detection

Before routing, we check for special composer modes that bypass standard routing:

Fusion โ†’ Multi-model DALL-E โ†’ Image gen Sora โ†’ Video gen Nano Banana โ†’ Photo-real images Deep Research โ†’ Perplexity
STEP 4

The Intelligent Router โ€” routeAI()

The heart of Yonoo. A 4-priority decision system that analyzes the question and picks the best AI:

Priority 0: Composer Mode Override โ–ถ
User explicitly chose a mode โ†’ route directly to that provider
Coding โ†’ Claude 4.5 (best at code)
Study โ†’ Claude 4.5 (educational)
Web Search โ†’ Perplexity (real-time)
Shopping โ†’ Perplexity (product search)
Reasoning โ†’ GPT-5.2 Pro (complex logic)
Priority 1: Content Analysis โ€” What is the question about? โ–ถ
Creative writing โ†’ Claude 4.5 Superior storytelling
Coding/programming โ†’ Claude 4.5 โ†’ GPT-5.2 fallback
Research/latest info โ†’ Perplexity Sonar Pro (live web search)
Analytical/reasoning โ†’ GPT-5.2 Pro (flagship reasoning)
General knowledge โ†’ GPT-5.2 (versatile)
Long context โ†’ Gemini 3 Flash (large context window)
Translation โ†’ Grok 3 (multilingual)
Casual/conversational โ†’ Llama 4 Scout (ultra-fast)
Priority 2: Task Type Context โ€” UI tile selection โ–ถ
Write / Create โ†’ Claude 4.5
Learn โ†’ Perplexity
Ask / Fix โ†’ GPT-5.2
Plan โ†’ Gemini
Translate โ†’ Grok
Kids โ†’ Claude (safe, engaging)
Priority 3: Balanced Round-Robin Fallback โ–ถ
If no clear signal โ†’ distribute across all available AIs using message hash
Ensures all 8 models get used, no single model is overloaded
Builds diverse usage data for future optimization

๐Ÿ”ฎ Yonoo Fusion

When one AI isn't enough. Multiple models answer the same question โ€” then the best insights are synthesized into one superior response.

User Question
โ†’
GPT-5.2
Claude 4.5
โ†’
Synthesis Engine
โ†’
Best Answer

Perplexity charges $40/mo for this. Yonoo includes it in every plan.

STEP 5

Perplexity Grounding Layer

For factual queries, Yonoo first fetches real-time context from Perplexity (web search), then feeds those facts to the routed AI. This means even Claude or GPT answers are grounded in current reality โ€” not just training data.

STEP 6

AI Generates Response

The selected model (or Fusion synthesis) generates the response. Each provider uses its native API format โ€” OpenAI-compatible for most, custom SDK for Claude. Yonoo handles all the API differences transparently.

STEP 7

Attribution & Delivery

Every response arrives with full transparency: which AI answered, why it was chosen, and any web sources used. The user always knows which model they're talking to โ€” that's the Yonoo promise.

Claude 4.5 Sonnet Coding
"Code/programming detected โ€” using Claude 4.5 for best coding assistance"

Yonoo vs. The Competition

FeatureYonoo Pro
โ‚ฌ19.99/mo
ChatGPT Plus
โ‚ฌ22/mo
Claude Pro
โ‚ฌ22/mo
Perplexity Max
$40/mo
GPT-5.2โœ“โœ“โœ—โœ“
Claude 4.5โœ“โœ—โœ“โœ“
Perplexity (live search)โœ“โœ—โœ—โœ“
Gemini, Grok, DeepSeek, Llama, GLMโœ“โœ—โœ—โœ—
Multi-model Fusionโœ“โœ—โœ—โœ“ ($40)
DALL-E + Sora + Seedanceโœ“DALL-E onlyโœ—โœ—
Intelligent auto-routingโœ“โœ—โœ—โœ—
Student plan (โ‚ฌ5.99)โœ“โœ—โœ—โœ—
Family plan (5 users)โœ“โœ—โœ—โœ—

Target Market: David's Advantage

๐ŸŽ“

Students (Primary)

50M+ European university students. Can't afford โ‚ฌ22/mo for ChatGPT Plus. Yonoo Student at โ‚ฌ5.99 gives them everything. Study modes, quizzes, citations โ€” built for learning.

๐Ÿ‘จโ€๐Ÿ‘ฉโ€๐Ÿ‘งโ€๐Ÿ‘ฆ

Families

Parents + kids sharing one AI subscription. โ‚ฌ29.99 for 5 users = โ‚ฌ6 each. No competitor offers this. Built-in Kids mode with safety guardrails.

๐Ÿ’ก

Small Wallets, Big Needs

People who want access to the best AI but can't justify โ‚ฌ22-40/mo per service. Yonoo democratizes access to frontier AI for everyone.

Pricing Architecture

Free ยท โ‚ฌ0 ยท 50 chats/wk
Student ยท โ‚ฌ5.99/mo
One ยท โ‚ฌ9.99/mo
Pro ยท โ‚ฌ19.99/mo
Family ยท โ‚ฌ29.99/mo (5 users)
Enterprise ยท Custom

All tiers get all 8 AI models. Limits are on volume (chats/week, images), not quality. The Spotify model: same music, pay for more access.

Technical Stack

Frontend โ–ถ
SPA โ€” Vanilla JavaScript, no framework (fast, lightweight)
PWA โ€” Install as app on any device, offline support
Chat UI โ€” Markdown rendering, code syntax highlighting
Composer Menu โ€” 16 specialized modes via "+" button
Backend โ–ถ
Node.js + Express โ€” Single unified server
PostgreSQL โ€” Users, subscriptions, analytics, waitlist
Stripe โ€” Payment processing, webhooks, subscription management
Replit Object Storage โ€” Universal file library
AI Providers (8 models) โ–ถ
OpenAI GPT-5.2, GPT-5.2 Pro, GPT-5.2 Instant, DALL-E 3, Sora
Anthropic Claude 4.5 Sonnet
Perplexity Sonar Pro (live web search)
Google Gemini 3 Flash + Nano Banana (images)
DeepSeek V3.2
xAI Grok 3
Meta/Groq Llama 4 Scout
Z.ai GLM-4.7
Media Generation โ–ถ
DALL-E 3 โ€” Artistic / creative images (OpenAI)
Nano Banana โ€” Photorealistic images (Google Gemini)
Sora โ€” AI video generation (OpenAI)
Seedance 2 โ€” Cinema-grade video with audio (ByteDance/fal.ai)
Lyria โ€” AI music creation (Google DeepMind)

Current Traction (February 2026 โ€” Honest Numbers)

501

Registered Users

Organic growth, no paid ads. 491 free + 4 Pro + 5 Family = 9 paying subscribers.

2,844

Chat Messages

Real conversations with real AI models. Growing daily.

21

Likely Flyer Scans

Physical QR flyers at TUM Munich. First offline marketing test.