The problem started when I was looking for a bot for Google Calendar a few months ago.
I realized that bot catalogs do not keep it up to date, therefore many bots are simply dead and don’t reply.
After my research based on 1,000 Telegram bots, I found that only 33.8% chatbots are online and ready to use, i.e. reply instantly, have filled descriptions, and are without adult content.
So I decided to create a proper catalog to give people the widest choice of bots, and with the best quality. To achieve these goals, I work hard with testing bots and their properties, like status, time of response, supported languages, and even category relevance.
BotoStore catalog uses many factors to find, test, and sort bots in order to show people the active ones first.
Q: Do I automize the testing chatbots or is it manual for now?
A: It was initially manual, but now BotoStore is automized with Human Augmented AI.
BotoStore’s bot 🤖 Scout can find and test chatbots through a public API. Another one — 🔎 Detector — detects the language of the bot, utilizes keywords for Categories, and filters out spam and adult content in order to restrict inappropriate items from the listings.
I’ve focused on 2 platforms for chatbots — Facebook Messenger and Telegram — and reviewed more than 50,000 chatbots. As a result, BotoStore has more than 25,000 chatbots; ordered by time of reply, the language of chat, and with more than 20+ parameters.
If BotoStore proves to be useful to its users, I will expand it to other platforms.