I built Mimir because I kept running into the same problem: deciding what to build, how to build it, and how to prioritize it against everything else.

Tools like Cursor make it easier to write code. But they don’t help you decide what code is worth writing. Mimir is my attempt to close that gap.

Here’s what it does:

- Paste or upload customer interviews, feedback, support tickets, usage notes (text, PDFs, screenshots)

- Extracts structured entities (pain points, feature requests, quotes, metrics)

- Synthesizes themes across all sources with evidence attribution

- Generates prioritized recommendations with impact projections

- Produces development-ready specs and pushes them to GitHub so agents (or humans) can start building

- Includes different chat modes for exploring ideas, refining scope, or generating new directions

The goal is to turn messy qualitative input into something structured, defensible, and ready for execution.