I used to do DevRel at Roblox and Rec Room. Way too much of my week went to copying feedback from Discord, forums, and support tickets into a spreadsheet. Then trying to figure out if the complaint on Discord was the same thing someone posted about on the forum three days ago. Built Chatter to automate that:

- Pulls feedback from Discourse, GitHub Issues, Discord, Webhooks, with more integrations on the way - Clusters related items using AI (batch classification on Haiku) - Can find community solutions and mark them for internal visibility - Custom scoring axes for prioritization if needed - Can email users when their reported issues get resolved

Tech stack: Node/Express, React, PostgreSQL, Railway, and Claude Haiku. I got classification costs to pennies a day with filtering and batching.

Demo (no login): https://demo.chatter.plus (Pulls data from the Godot engine community (Github Issues and Forums) every morning)

https://chatter.plus

Really excited to start sharing this wider. I started with a much bigger idea a few months ago and feel like I've honed it to a clean and lean offering that just needs some tweaking.

Curious how others handle this? This is definitely a problem for projects with big followings, but I think people would be surprised by how fast feedback can get out of hand.