Skip to main content
Mythos

Voice Input in 📝MythOS is part of 📝Editor Core: hold a key, speak, and get structured text back — transcribed, then run through your own abbreviations and auto-linked to the memos and tags you already use.

Dictation elsewhere gives you a wall of plain prose that you then have to format by hand, which costs more than typing would have. Speech is only faster if what lands in the memo is already marked up.

It is for anyone who thinks out loud faster than they type — capturing a journal entry on a walk, or drafting into a memo without touching the keyboard.

Key Capabilities

  • Hold or latchCmd/Ctrl + Shift + D held dictates while you speak; double-press the same keys to latch recording on.
  • Whisper transcription — audio goes to OpenAI Whisper and comes back as text, with a 60-second cap per take.
  • Dictation Dictionary — your own triggers expand on transcription, matched case-insensitively and longest-first.
  • Four expansion types — a trigger can expand to an abbreviation, a @mention, a hashtag, or plain text.
  • Spoken hashtags — saying "hashtag focus" produces a real tag chip rather than the literal words.
  • Automatic mentions and tags — memo titles and tags you already use are linked in the transcript without being asked for.
  • Microphone picker — choose an input device, with a live waveform while recording.

Getting Started

  1. Add your own OpenAI key under Settings → AI — voice prompts you to set one up on first use.
  2. Open any memo or daily memo and press Cmd/Ctrl + Shift + D, or use the floating mic button.
  3. Hold the keys while speaking, or double-press to latch, then stop when you are done.
  4. Build a Dictation Dictionary so your shorthand expands the way you want it to.

FAQ

  • Do I need my own API key? Yes — voice runs on your own OpenAI key, and the mic prompts you to add one before it will record.
  • How long can one take be? Sixty seconds, after which recording stops on its own. Dictate in passes rather than monologues.
  • What if it hears a memo title? It links it. Any title in your own library, and any tag you already use, is converted into a real mention or tag chip.
  • Will it link something I did not mean? It can. Matching is by whole word against every title and tag you own, so a common-word title gets picked up — worth a read-through before saving.
  • What language does it transcribe? English. The transcription request is pinned to English rather than auto-detecting.
  • Where does it work? Any memo, the daily memo, and the floating mic available across your library.

Related

  • 📝Mention — the @ typeahead that voice writes into your text automatically
  • 📝Tagging — the hashtag system spoken tags feed into
  • 📝BYOK — bringing your own model key, which voice depends on

Contexts

Created with 💜 by One Inc | Copyright 2026