Mimir

hey hackers - i built this tool called mimir because i kept having this problem where i'd do a bunch of customer interviews or collect feedback and then just...not know what to build first.

basically its a way to take all your messy data - like raw customer interviews, support tickets, feature requests, usage metrics, whatever - and synthesize it into actual recommendations of what to build next, with projected impact on key business metrics.

for example you can paste in feedback from users, upload sales call notes, screenshots of reviews, etc. and it clusters everything into themes and tells you "8 people mentioned this, 2 people mentioned that, you should probably build the first thing" all powered by ai (claude) so you can chat with it to refine the recommendations or ask why it thinks something is high impact. then it connects to github and can actually generate dev specs from the recommendations...so if you have an agent connected to github issues, it can actually build it right away.

i built it mostly for myself because i was tired of reading through notion docs trying to figure out what mattered, but figured other founders might have the same problem!

anyway would love to know what you guys think. like is this actually a problem you deal with or am i the only one who sucks at synthesis.