Technology & Infrastructure
Built for traders,
everywhere.
A global integrated trading education and technology ecosystem — launched from India, combining structured educator-led learning, TradingView Pine Script strategies, MetaTrader 5 automation, and AI-native operations in a single, scalable platform.

Platform Architecture
Four layers.
One ecosystem.
SmartifyTrade is designed so every layer — education, execution, intelligence, and infrastructure — is independently replaceable, scalable, and cloud-agnostic. No lock-in to a single provider, model, or hosting platform. A change in any one layer (a new cloud region, a new AI provider, a new broker integration) never requires touching the others, which is what lets the platform grow without a rebuild.

Education Layer
- ◆Educator Dashboard
- ◆Student Dashboard
- ◆Course Publishing
- ◆Live Class Scheduler
- ◆STEP Workspace
Execution Layer
- ◆TradingView Webhook Intake
- ◆Signal Router (S1 · Europe)
- ◆License Validation Engine
- ◆MetaTrader 5 Bridge EA
- ◆Signal Queue + ACK Lifecycle
AI Layer (Limitify AI)
- ◆Kuku Chatbot (S2)
- ◆Educator Co-Pilot
- ◆Conversion Intelligence
- ◆Aakruti Scanner
- ◆FOPS Gateway (Port 7771)
Data Layer
- ◆Firebase Auth (Google + Phone)
- ◆Firestore (real-time NoSQL)
- ◆SQLite (License + Signal DB on S1)
- ◆Learning Logs (LRBI)
- ◆Analytics Snapshots
Signal Flow — TradingView to Live Execution
TradingView Alert → Webhook → S1 Signal Router → Signal Queue
↓
MT5 Execution ← Bridge EA ← EA Pull (every 1s)
↓
ACK + Result LogLimitify AI — Embedded Intelligence Layer
AI is not a feature.
It is the platform's operating layer.
Six production AI agents power SmartifyTrade — master control, student support, platform intelligence, educator screening, curriculum tooling, and a multi-matter portal engine. Agents run on a multi-provider stack — Anthropic Claude and OpenAI — routed per task, with additional providers ready (Google Gemini, local Ollama). All learning logged. All decisions auditable.

Limitify AI
builds and manages the agents
SmartifyTrade
hires those agents (Client #1)
14
Agents built
11
Live
3
Planned
FPA
LIVEFounder Personal Agent · Master Control
Master control agent for the entire platform. Interprets founder intent, analyzes risk/impact of every proposed action, and coordinates execution across all other agents. Three operating modes (Normal / Review / Emergency), zero autonomous production writes — every state change requires explicit founder approval.
Kuku
LIVEStudent Support Chatbot
24/7 AI support agent embedded on the website. Handles student queries about the platform, trading education, EA setup, and account issues in English, Hindi, and Hinglish. Powered by a multi-provider stack with automatic escalation to a structured ticket system on low confidence.
Aakruti
LIVEPlatform Intelligence Scanner
Platform-wide anomaly detector. Watches signal execution, user behaviour, agent performance, and system health. Reports to FPA dashboard — never directly to users. Approval-gated by founder before any action. Three trigger modes: scheduled batch, pure-regex log scan (zero AI cost), on-demand.
Educator Application Scanner
LIVEScreening & Verification
Screens inbound educator applications against platform policy — checks YouTube presence, content history, structural quality, and produces a structured scan report for the founder to review. Read + report only. No authority to approve, reject, or contact an applicant directly.
Educator Co-Pilot
LIVECurriculum AI
AI assistant inside the Educator Dashboard. Generates course structures, identifies content gaps, imports YouTube videos into knowledge graphs, and helps educators build teaching pathways faster. Two phases live (Readiness Scan for prospects, Deep Channel Scan for onboarded educators); Knowledge Graph phase planned.
Matter AI
LIVEMulti-Matter Portal Engine
One engine, multiple "matters" — each with its own knowledge folder, system prompt, user whitelist, and escalation path. Powers the public AP recruitment portal today, with private OTP-gated matters (legal, CA) queued behind the same engine. Shared-secret gate, per-IP rate limits, and a 4-tier document visibility model — every escalation routes through human review.
Every AI agent has an LRBI Loop — Log · Rate · Bug · Improve — building a permanent learning history and performance track record from Day 1.
PlannedVoice-assisted learning — an AI-narrated tutorial library (platform, strategy, indicator, and EA walkthroughs, hosted on our YouTube channel) and a voice mode for Kuku, so support and onboarding work by voice as well as text.
Technology Stack
Modern. Proven.
Replaceable.
Every component in the stack — hosting, database, authentication, CDN, charting, and execution — was chosen for scale, developer velocity, and portability, not because it locks in the fastest. If a provider changes pricing, deprecates a feature, or has an outage, the adapter pattern underneath means swapping it out is a configuration change, not a rewrite.

Frontend
Next.js 14
App Router, Server Components, Edge functions
Language
TypeScript
Strict-mode type safety across entire codebase
Auth + DB
Firebase
Authentication, Firestore, Cloud Functions
Deployment
Vercel
Auto-deploy, edge network, preview environments
AI Providers
Claude + OpenAI
Multi-provider: Anthropic Claude + OpenAI GPT, per-task routing
Backend
Python
Signal router, AI agents, background jobs
Execution
MetaTrader 5
Broker connectivity for automated trading
Signals
TradingView
Pine Script strategies + webhook alerts
Database
Firestore
Real-time, globally distributed NoSQL
Runtime
Node.js
API routes, server-side logic
Version Ctrl
GitHub
Source control, CI/CD, code review
CDN + DNS
Cloudflare
Global CDN, DDoS protection, DNS
Design System
Tailwind CSS
Utility-first styling, consistent design tokens across every page
Web Server
Nginx
S2 reverse proxy — routes chatbot-api, aakruti-api, and FPA API relay
Strategy Lang
Pine Script
TradingView-native language for the 12 SmartifyTrade strategy frameworks
Scale & Security
Enterprise-grade
from Day 1.
Multi-server failover, HMAC-signed auth, OTP-gated critical actions, zero secrets in code. Built to protect live trading accounts and educator revenue from the first user, not retrofitted after a scare — because a platform executing real trades on real broker accounts doesn't get a grace period to add security later.

Target Users
100,000+
active accounts
EA Installations
10,000+
projected Phase 1
Signal Latency
~1-2s
TradingView → broker
EA Polling Interval
1 sec
per account
Multi-Cloud Ready
3 clouds
AWS · GCP · Azure
AI Providers
4 ready
Anthropic · OpenAI · Gemini · Ollama
◆HMAC-Signed Auth Tokens
FPA and apply verification use HMAC-signed tokens with short TTLs. No session storage on server.
◆OTP-Gated Critical Actions
Destructive FPA operations require TOTP verification — not just password. Educator approvals, role changes, platform flags.
◆Server-Side Application Routes
All apply flows (Educator, GP, AP) go through /api/apply/* — never direct client Firestore writes. Every submission validates AP code server-side.
◆Rate Limiting on Public Routes
IP-based rate limiting on all public API endpoints. Firebase Security Rules mirror API auth requirements.
◆Multi-Server Failover
S1 EA router (Europe) has automatic failover to S2 (Asia) if primary goes down. Zero-touch recovery.
◆Zero Secrets in Code
All API keys, tokens, and secrets in environment variables. .gitignore enforced. Automated secret scanning on commit.
Engineering Philosophy
Every dependency,
replaceable in one day.
Cloud provider outage? AI model deprecation? Payment processor issue? Every external service goes through a gateway pattern with a backup adapter ready. Switchover time is measured in hours, not weeks.

| Service | Current | Backup Ready | Switchover |
|---|---|---|---|
| AI / LLM | Anthropic Claude + OpenAI | Google Gemini · Ollama | 4 hours |
| Cloud Hosting | Vercel + Multi-Region VPS | Railway · AWS · GCP | 2 hours |
| Database | Firebase Firestore | Supabase · PlanetScale | 1 day |
| Payments | Razorpay (planned) | Cashfree · PayU · Stripe | 4 hours |
| CDN | Vercel Edge | Cloudflare · Bunny.net | 2 hours |
| SendGrid (planned) | AWS SES · Postmark | 2 hours | |
| SMS / OTP | Firebase Phone Auth | MSG91 · Twilio | 4 hours |
| File Storage | Cloudflare R2 (planned) | AWS S3 · Backblaze B2 | 4 hours |
Company Recognition
Registered. Recognised. Built to last.
Legal Entity
SmartifyTrade Technologies
Private Limited
Corporate Identification (CIN)
U85500UW2026PTC254424
Recognition
DPIIT Recognised Startup
Under Startup India, Ministry of Commerce & Industry
Registered Office
A-116, Sec-132, URBTECH Trade Centre,
Noida, Gautam Buddha Nagar,
Uttar Pradesh 201304, India
Domain
Stage
Pre-launch, active development.
Payments launch: August 2026.
Explore the platform.
See how SmartifyTrade combines education, execution, and AI in one integrated ecosystem.