SuperWhisper
Cloud + LocalCloud-first with local fallback. Powerful LLM postprocessing, but context data leaves the device.
Independent benchmarks — WER, CER, and PER on identical audio samples. No vendor ties, no paid placements.
See overall rankingsCloud-first with local fallback. Powerful LLM postprocessing, but context data leaves the device.
Best-in-class transcription accuracy. Cloud-only; everything uploads to their servers.
Excellent local transcription for Mac. File-based workflow, no real-time dictation.
Our methodology
We test every app on the same ElevenLabs-synthesised audio fixtures under identical conditions. Word Error Rate, Character Error Rate, and Punctuation Error Rate are computed against multi-reference ground truth. All test data and diffs are published.
Read the methodology →Different jobs need different tradeoffs — speed, privacy, accuracy, price.