Documentation

Everything you need to dictate.

One page, four sections. Installation, first dictation, custom dictionary, and how the trial & license work.

1. Installation

Requirements

  • macOS 14 Sonoma or later.
  • Apple Silicon (M1 or later). nommo runs speech recognition on the Neural Engine and the cleanup model on the GPU, both on-device. Intel Macs are not supported.
  • 8 GB RAM is fine. nommo peaks around 1.3 GB of memory and settles near 600 MB between dictations.
  • About 800 MB of disk for model weights, downloaded once on first launch. Everything runs offline after that.

Install the .dmg

  1. Download nommo from /download and open the .dmg.
  2. Drag nommo to your Applications folder.
  3. Launch nommo from Applications. On first launch macOS verifies the signature; this can take a few seconds.
  4. On first launch nommo downloads the speech and cleanup models (about 800 MB). Subsequent launches are instant and fully offline.

Grant permissions

nommo asks for two permissions when it starts. Both are required.

  • Microphone: needed to capture your voice. Audio is processed locally and never leaves your Mac.
  • Accessibility: needed to place transcribed text into the app you are dictating into. nommo never reads the contents of other apps — it only uses Accessibility to type text forward.

If you accidentally deny either permission, open System Settings → Privacy & Security → Microphone / Accessibility, and toggle nommo on.

Updating from 1.x

nommo 2.0 downloads the new speech and cleanup models once (about 800 MB) the first time it launches. The old models aren’t removed automatically — free about 4.6 GB by going to menu bar icon → Settings → Advanced → Remove old model caches.

Running macOS 13 Ventura? nommo 2.0 requires macOS 14 Sonoma or later, so it won’t install on your Mac. Version 1.2.1 stays available at github.com/Nommo-AI/releases (v1.2.1).

2. First dictation

nommo lives in your menu bar. Its icon appears as soon as you launch the app — there is no separate process to start up. If models are still downloading, progress shows in the Home window and in the menu bar icon.

Default hotkey

Hold Right Option to record, release to transcribe. A small pill appears near the bottom of the screen while you are speaking. Release the key and your polished text lands in whatever app has focus — email draft, Slack thread, code editor, Notes.

Push-to-talk vs. toggle

Two recording modes ship in V1:

  • Push-to-talk (default): hold the hotkey to record. Release to stop. Best for short dictations and conversational speech.
  • Toggle: press once to start recording, press again to stop. Better for longer monologues where you do not want to hold the key.

Switch modes from the menu bar icon → Recording Mode.

What the pipeline does

Every dictation runs through three local stages: speech recognition on the Neural Engine, a cleanup model for filler removal and grammar cleanup at your chosen intensity (Off, Standard, or Polished — Standard is the default, in Settings > General), and your custom dictionary for domain substitutions. Short phrases land in roughly 100 ms after you release the hotkey on a recent M-series Mac.

3. Custom dictionary

The custom dictionary teaches nommo names, jargon, and technical terms specific to your work. Your entries feed a phonetic corrector that runs right after speech recognition, plus an exact-match pass after cleanup that restores your spelling. Add a “Sounds like” pronunciation hint for any term nommo tends to mishear.

Add a term

  1. Click the nommo menu bar icon → Dictionary.
  2. Click Add term, type the word or phrase as you want it to appear, and confirm.
  3. Terms are stored locally in a SQLite database at ~/.nommo/captures.db.

What works well here

  • Proper names — colleagues, clients, partners.
  • Company / product names that speech recognition would otherwise spell phonetically.
  • Industry acronyms (ISO, HIPAA, SOC 2, LLM).
  • Repeating technical vocabulary in your field.

Every term in your dictionary is checked with phonetic matching on every dictation, so there is no cap on how many terms nommo checks.

4. Trial & license

14-day free trial

Download nommo and you get every feature for 14 days — the full cleanup pipeline, all nine app categories, the custom dictionary, push-to-talk and toggle modes. No credit card, no account. You just start dictating.

A small banner in the menu bar icon shows the days remaining. When the trial ends nommo stops transcribing new dictations until a license is activated.

Buy and activate

  1. Buy nommo Pro ($99 one-time) from the pricing section. Checkout is handled by Lemon Squeezy — they are the merchant of record.
  2. After purchase you receive a license key by email. Keep it somewhere safe.
  3. In nommo: menu bar icon → License → Activate, and paste the key.
  4. The key activates on your Mac. A single license covers up to 3 Macs you personally use.

License transfer

Getting a new Mac? Deactivate the old one first (menu bar → License → Deactivate), then activate the new one with the same key. You can deactivate from any Mac at any time — no support ticket needed.

What Pro unlocks

  • Unlimited dictation beyond the 14-day trial.
  • License transfer across your own machines (up to 3).
  • Priority email support at support@getnommo.com.
  • Early access to new features.

Updates within 1.x and 2.x are included free. Future major version upgrades (V3 and beyond) may be separate purchases at a reduced upgrade price. We will never take features away from a version you paid for.