3
4 Comments

I made an AI-powered chat moderation bot for your Telegram community

Hey IH makers!

👋 I'm Ben and I recently built https://watchdog.chat, an AI-powered chat moderation bot for your Telegram group. Product Hunt link: https://www.producthunt.com/posts/watchdog-4

🍿 Demo video:
https://www.youtube.com/watch?v=VDQd9DFbTz4

🤖 Fully automate chat moderation for your Telegram community with AI and never worry about rulebreakers, trolls, or spambots again!

👮‍♂️ Define rules in plain English. You can even customize the rules and associated punishments per community or per channel

💰 Pricing starting at just $9/mo, with an additional 2 months free if you pay annually. Free trial available with all plans.

🤮 Allergic to subscriptions? Don't worry, there's also a one-time (credit-based) pricing if your community is just getting started, and the credits never expire 🙌

🙏 Check it out and let me know what you think! Thanks so much.

posted to Icon for group Building in Public
Building in Public
on May 4, 2024
  1. 3

    I'm one of the early beta testers for Watchdog. It's awesome to witness the potential for AI in moderating trolls and undesirable messages in my communities and chat groups. Since adding Watchdog to my local chat group, I no longer worry about not catching spam messages in the chat while I'm asleep.

    1. 1

      Thanks Jason! Has been great working with you.

  2. 1

    So it's kind of a "Dyno" but for telegram?

    1. 1

      It's similar, but better. Dyno isn't able to intelligently monitor messages, it just looks for banned words which leads to too many false positives in my experience.

      If your banned word is "fuck" for example, you wouldn't want to ban someone that says "hey, that's really fucking cool!" or "fuck, I'm sorry to hear that man" but you probably would if they tell another member to "go fuck themselves."

      That's where Watchdog shines. You can define rules in plain English like "no insulting language towards other members" instead of simple banned words/string filters and the number of false positives is significantly reduced that way.

      I'll add some "compared to" pages on the landing page to make the differences between existing solutions and Watchdog much more clear.

      Thanks for the comment!

Trending on Indie Hackers
Stop Spamming Reddit for MRR. It’s Killing Your Brand (You need Claude Code for BuildInPublic instead) User Avatar 193 comments What happened after my AI contract tool post got 70+ comments User Avatar 154 comments Where is your revenue quietly disappearing? User Avatar 58 comments How to build a quick and dirty prototype to validate your idea User Avatar 53 comments The Quiet Positioning Trick Small Products Use to Beat Bigger Ones User Avatar 40 comments I Thought AI Made Me Faster. My Metrics Disagreed. User Avatar 38 comments