Lispr

Most voice tools make you open a chat window, speak, copy the result, switch back to your app, and paste. Lispr removes every step after ‘speak.’ Hold the right Option key to dictate; add Control to translate. Text inserts at the active cursor — in Slack, Notes, Figma, Terminal, or anywhere else a cursor blinks. The vendor describes 34 supported languages detected automatically, no menu required, and clipboard state is preserved so whatever you had copied before is still there after.

The differentiating detail is the translation path. Most Mac dictation tools — including Apple’s built-in option — have no translation layer at all. Lispr treats translation as a first-class gesture, not a post-processing step. The vendor states the model is Whisper large-v3 over an encrypted connection, audio is discarded after transcription, and nothing is retained or used for model training. No account exists to breach.

This fits a narrow but real scenario well: a bilingual professional who types the same phrases in two languages repeatedly and wants the friction gone without paying for a subscription or running a heavier application. It does not fit teams who need to embed voice input in a product, trigger downstream actions, or process audio in batch. The comparison table on the vendor page positions it against Wispr Flow — which the vendor describes as a full app with a subscription — and Apple Dictation, which has no translation. Lispr’s footprint is approximately 4 MB as a menu bar item with no persistent window.

The app is free during early access with no account or subscription required. The vendor states it is notarized by Apple and built by Codebridge Technology, Inc., a named company. No API is available and no self-hosted option exists, so the audio-to-transcription path goes through Codebridge’s infrastructure regardless of preference.