Blog

Notes on dictation, speech recognition and voice typing

Notes from building an on-device dictation app: what speech models get wrong, why "cleaning up" is a harder promise than rewriting, and what runs on your machine versus someone else’s.

Guides

  • Dictating prompts to ChatGPT and Claude

    Why spoken prompts come out longer and more specific than typed ones, and how to dictate them into a browser chat box on a Mac without losing the text.

  • Writing issues an agent can act on

    An issue is a prompt that outlives the conversation. How to dictate issues, tasks and review comments so an agent — or a person — can act on them.

  • Voice input for Claude Code

    How Claude Code’s built-in /voice relates to a system-wide dictation layer, and how to dictate prompts in the terminal, Cursor and the browser with Voicecape.

Articles

  • What Voicecape sends over the network

    Three requests, all named and verifiable from outside the app: a licence activation, a background re-check of it at most weekly, and a daily update check.

  • Dictation for developers

    Code is a bad thing to dictate. The prose around code — commit messages, PR descriptions, review comments, docs — is a very good one.

  • macOS permissions for dictation

    Microphone, Accessibility and Input Monitoring: what each is for, why macOS asks separately, and what to do when one looks granted but nothing happens.

  • Typing with your voice in any app

    How text gets from a dictation app into the window you are looking at on macOS: accessibility APIs, the clipboard path, and why browsers differ.

  • Korean voice typing on a Mac

    What makes Korean harder for a speech model than English, and why filler removal works differently in a language without spaces between clauses.

  • One model, five languages

    How a single multilingual speech model handles English, Korean, Spanish, Japanese and Chinese — including a sentence that switches language halfway through.

  • Choosing a Whisper model size

    Why a bigger speech model is not simply better, what changes between tiny, base, small and large-v3-turbo, and the failure mode that made us ship a 547 MB model.

  • Why Mac dictation gets names wrong

    Speech models predict likely words, and your colleague’s name is not a likely word. Why a personal dictionary fixes what a bigger model cannot.

  • Verbatim versus cleaned dictation

    When you want every word you said, and when you want the sentence you meant. Two modes, and how to tell which one a piece of writing needs.

  • What "cleaned, never rewritten" removes

    The exact list: fillers, accidental repeats, trailing filler phrases and orphaned commas. And the list of things deliberately left alone, including self-corrections.

  • Does dictation software send your voice to the cloud?

    How to find out for yourself, on any dictation app, in about five minutes — with airplane mode and a firewall monitor, rather than by reading a privacy policy.

  • On-device versus cloud speech recognition

    What differs between recognition that runs on your Mac and recognition that runs on a server: latency, cost, offline behaviour and who holds the audio.

  • Why dictation software rewrites what you said

    Most AI dictation tools do not transcribe you — they paraphrase you. Here is the technical reason why, and what a tool has to give up to stop doing it.

Back to voicecape.com