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