Neural-network PID tuner for Betaflight FPV drones. Pilots upload a blackbox log (.BBL/.BFL); the model runs spectral analysis on the gyro trace, flags resonance and noise frequencies, measures step response, and returns optimized PID, filter, and feedforward values as ready-to-paste CLI commands. Trained on freestyle, racing, cinematic, and long-range logs, with motor thermal prediction to avoid settings that would overheat motors. Supports Betaflight 4.3 through 4.5+.
FPVtune is newly emerging in Data & Analytics — it's early-stage, which means lower adoption risk but also less community validation.