Hi Product Hunt. I built Lidolovets after watching small service businesses lose deals for one boring reason: somebody in a group chat asked "can anyone recommend a cleaning company" and nobody saw it for two days. The hard part was never finding keyword matches. It was telling a buying message apart from everything else that contains the same words. "Looking for a cleaning company" and "the cleaning company ruined my sofa" are both matches and only one is a customer. So the product reads the open chats, a model decides whether the message is a real request, and you get the message itself with the author handle, not a score. Two rules I kept: open chats only, nothing closed or scraped, and the lead has to arrive in minutes, because in chats a two-hour-old request is already answered by someone else. It currently runs for Russian-speaking Telegram across 25+ niches. Happy to answer anything about the filtering, the false-positive rate or where this approach breaks.

Lidolovets

AI finds your clients in open Telegram chats

AI finds your clients in open Telegram chats