It started from wanting to manage notes and Telegram messages scattered across places in one spot, and have AI analyze/summarize them and flag what to handle that day.
← All work
Case 09 / 09
MemoHub
A web app that gathers scattered notes and Telegram messages, analyzes/summarizes them with AI, and even gives same-day reminders — deployed
Deployed (private)Idea→shiprecord incomplete*
2026.04 / Solo (AI as the team) / Web
- Problem
- Notes and Telegram messages were scattered across places.
- My role
- Planning · setting the bar · directing AI to build · deployment
- Stack
- Web · Vercel · Supabase (sync) · Groq API (llama-3.3-70b-versatile) · Telegram
Results
- Running on Vercel (live)
- Deployed
- Running on Vercel (live)
- AI analysis/summary (llama-3.3-70b)
- Groq
- AI analysis/summary (llama-3.3-70b)
01
Idea / Problem
02
The Plan & Bar I Set + Directing AI
- Planned the flow — collect messages → AI analysis/summary → same-day reminders — and directed AI step by step
- Wired the AI features with Groq (llama-3.3-70b) so analysis/summary actually run, and deployed on Vercel
03
Results
- Deployed live on Vercel, with Groq-based AI analysis/summary working
- A small, complete example of "build fast and get it deployed"
Next project →
THE HOUSE
Starting with zero game-engine experience, a first-person escape-room game built solo by directing AI — all the way to store submission