Player not loading? Watch on YouTube
The presenter explains why he moved away from Wispr Flow and uses TypeWhisper for dictation and file transcription. He describes it as open source and free for personal use, with separate licensing considerations for commercial use. The stated platforms are Mac and Windows, plus an alpha iPhone app.
The Mac setup covers microphone and accessibility permissions, speech model selection and keyboard shortcuts. Hybrid mode supports a short press to toggle recording or holding the key for push-to-talk. The presenter transcribes audio extracted from a roughly 20-minute video in under a minute on his setup and shows SRT and VTT export. That timing is a demonstration result, not a hardware-independent benchmark.
Dictionary corrections address misheard terms, while snippets expand spoken keywords into saved text. A cleanup workflow removes fillers and fixes grammar. The presenter explains how he adjusted its prompt after the model answered dictated questions instead of editing them. He also uses a cloud provider on an older Intel computer, so privacy depends on the selected processing route.
The comparison covers Gemini's macOS voice features and briefly shows subscribed ChatGPT/Codex dictation settings. The presenter disables Gemini's reasoning option because it interprets requests intended for his coding tool. He distinguishes cloud processing from TypeWhisper's ability to keep dictation local and private.
