Building TGLeads surfaced a new need: finding promising Telegram channels with potential leads for current and new clients.
Built large-scale parsers plus AI to catalog roughly 500,000 CIS Telegram channels, scored on WAU/DAU/MAU.
Identified core discussion topics and split channels into 45 categories and 10–15 subcategories, with a detailed map of "hot problems" across 10 categories.
Built a mechanism linking a user's replies across channels into one chain, to analyze personal and psychological traits.
The hard engineering behind the result.
Scoring and categorizing half a million channels accurately and keeping it current.
Linking replies into coherent per-user chains at that scale.
A look at what we shipped.
A happy customer.
The data let the client grow from 2 to 6 client-acquisition channels and more than double annual revenue.
More from Social Media & Networking
All work →DIVO Global
Telegram-powered relaunch of a fashion & modeling network, with AI face search.
Clicker Miniapp
A viral Telegram clicker game with referral rewards, built to drive an ICO token launch.
iLinz
High-fidelity online try-on for real eyewear frames, built to lift lens and glasses sales.