β 100% local inference β Wispr-optimized model runs on your hardware
β Model selection β tiny (75MB, CPU) to large-v3 (3GB, CUDA)
β Correction dictionary β teach your vocabulary, corrections
apply retroactively to history
β Toggle and push-to-talk modes
β Full transcript history w/ search and CSV export
β Compact floating widget (95px collapsed)
β CUDA GPU acceleration auto-detected
β Cross-platform β Linux, macOS, Windows
Free demo available. Full version: CA$29.99 one-time. No subscription.