Alpha

A thought spoken.
A sentence written.

Turn microphone audio into text, then bring it into the app where you’re writing. Speech recognition runs on your Mac.

Alpha 0.2.0 · Apple Silicon · macOS 15.0+

Hodani DictationMac

Let’s move the design review to Thursday. I’ll bring the updated screenshots.

Illustrative interface · sample content, not an app screenshot

Inside Dictation.

Use a press-to-toggle or hold-to-speak shortcut. Stop recording when your thought is ready.

Saved audio and transcript history let you listen, copy and reprocess. If insertion is unsupported or focus changes, your text remains available.

Keep the original language or use optional local translation. Translation needs macOS 26 or later and installed language assets.

What you’ll need: Apple Silicon · macOS 15.0 or later. Download the Parakeet speech model before first use. Microphone access records audio; Accessibility access inserts text. Optional translation needs macOS 26 and language assets.
Current boundaries: Recognition can make mistakes, especially with mixed-language speech. Review the result. Insertion depends on the destination app and field; secure fields are excluded. Real-world microphone, shortcut and editor compatibility testing remains part of beta preparation.
Privacy & connections: Speech recognition and supported translation run on your Mac, with no cloud inference fallback. Preparing models and language assets requires downloads. Saved audio and text stay in the app’s local library.

Every Hodani app is still in development. There are no final releases yet.

Alpha — Tray, Dictation, Term, Quicktools, Calls and Notebooks. Early builds; features and compatibility may be incomplete.

Beta — Pulse, Capture and Notes. Available to try and still being refined.

Downloads are for Apple Silicon Macs running macOS 15 or later. Keep backups of important work.