Fundedologist: trading academy, journal and AI coach in one platform

The challenge

Fundedologist sells no signals and no funded accounts. What it promises is discipline, and discipline has to show up in a trader's own data. The product had to teach first, then pull real trades in automatically, then give feedback grounded in those numbers, for Persian readers right-to-left first and English readers second.

What we did

We designed and built the whole system as one Next.js 16 application on Vercel, with Supabase for data, Clerk for accounts and Sanity for articles. The public site runs fully right-to-left in Persian, with hreflang, structured data, a live sitemap and an llms.txt file for AI search. Inside the dashboard, custom MetaTrader 4 and 5 Expert Advisors sync closed trades on their own, and CSV import covers everything else. The AI coach first computes a fixed performance snapshot and rule-based behaviour flags such as revenge trading, overtrading and broken plans, and only then asks Claude for a structured review, so the advice cannot invent numbers. A prop-challenge simulator, session replay on historical market data, daily missions and a 1,000-point discipline score turn practice into habits, while web push and a Telegram bot keep traders in the loop.

What we built

• Automatic trade sync from MetaTrader 4 and 5, plus CSV import
• An AI coach that reviews a computed snapshot, not raw guesses
• Prop-challenge simulator and replay on historical candles
• Persian right-to-left first, English second
• Missions, XP and a 1,000-point discipline score
• Notifications by web push and a Telegram bot

The outcome

Fundedologist is live in Persian and English with an eight-course, 62-lesson academy, three plans starting with a free tier, and a dashboard that installs as a progressive web app. Every public page is server-rendered with hreflang and structured data, and the discipline score deliberately gives zero points for profit.