I got tired of
heading in blind.

Every competitor knows the night-before ritual. You're watching footage for the third time, trying to figure out what that opponent does from half guard, scribbling notes into your phone, telling yourself you're prepared. You're not. You're just tired.

The problem isn't that the footage doesn't exist. Competition streams, YouTube matches, uploaded clips — there's more BJJ footage available now than at any point in the sport's history. The problem is that watching video without a framework is just watching video. You notice the same things you already knew. You miss the patterns that would actually change how you prepare.

The athletes who've solved this have a coach watching their footage, building a breakdown, writing a gameplan. That's how the top competitors prep. But most of us don't have that — we're self-coached, or our coach has thirty other students, or we just don't have the budget for someone to sit with our footage for two hours before every tournament.

The tools that existed were either built for teams (Dartfish costs what some people pay in rent, and assumes you have a dedicated analyst) or completely generic (asking a general AI to break down a BJJ match doesn't work — it doesn't know half guard from side control). Nothing was built for the individual competitor who just wants to be better prepared than they were last time.

So I built it myself. Paste a link to a match — yours or your opponent's — and the AI does what a good analyst does: identifies every position, every key moment, the tendencies that keep showing up. For tournaments, connect your Smoothcomp bracket and it scans every opponent in your draw and writes you a gameplan for each of them. On competition morning, one screen shows you exactly what to open with and what danger to watch for. No reading required.

RollPlan is still in beta. The AI isn't perfect — position labels can be wrong, and sometimes it misses a detail a human coach would catch. Every incorrect label can be corrected inline, and every correction makes future analysis better. The mobile app isn't out yet. There are rough edges.

But the core thing works: you upload footage, you get a real breakdown, and you go into your next competition knowing more than you did. That's what this is about.