Repolode
google-gemini

google-gemini/jot-gemini-transcribe-macOS

A macOS menu bar dictation app: hold fn, speak, and polished text is typed at your cursor using Gemini transcription.

Open-source alternative to:Apple DictationWispr Flow

Stars
555
Forks
58
Open issues
12
Last commit
2 months ago

What is it?

Jot is a macOS dictation app that records while you hold the fn key and inserts the transcribed text where your cursor is. It removes filler words, adds punctuation, and applies self-corrections, so saying "let's meet at 1pm, actually, make it 2pm" produces "Let's meet at 2pm." Audio is sent directly from your Mac to the Gemini API using your own API key, with no intermediary server or account.

The app is built around reliability: audio is written to disk from the first moment of recording so crashes don't lose speech, offline dictations queue until the connection returns, and every failed transcription can be retried from the History window. A Dictionary feature passes custom names and terms to the model alongside the audio. Insertion tries the Accessibility API first, then a guarded paste that restores the clipboard.

The project is written in Swift and requires macOS 14 or later and Xcode 16 or later to build. The core logic lives in a headless Swift package, JotCore, which has its own tests. It is an unofficial demo and is not officially supported by Google.

Key features

  • 01Hold fn to record and release to insert text at the cursor.
  • 02Hands-free mode keeps recording after the key is released.
  • 03Removes filler words and handles mid-sentence corrections.
  • 04Audio is saved to disk during recording and recovered after crashes.
  • 05Custom dictionary terms are sent with the audio to improve recognition.
  • 06Transcription history with retry for failed dictations.

Use cases

  • Drafting emails, chat messages, or documents by speaking instead of typing.
  • Dictating code-related names and technical terms with a custom dictionary.
  • Capturing notes quickly on a Mac without switching windows.

Quick start

  1. 1

    Download the latest Jot DMG from the Releases page

  2. 2

    Drag Jot into Applications and launch it from there

  3. 3

    Paste a Gemini API key from Google AI Studio and grant microphone and Accessibility permissions

Things to know

  • Runs only on macOS.
  • Requires your own Gemini API key, and you pay Google for usage beyond the free tier.
  • This is not an officially supported Google product.

In collections

Topics

FAQ

What is jot-gemini-transcribe-macOS?
A macOS menu bar dictation app: hold fn, speak, and polished text is typed at your cursor using Gemini transcription. Jot is a macOS dictation app that records while you hold the fn key and inserts the transcribed text where your cursor is. It removes filler words, adds punctuation, and applies self-corrections, so saying "let's meet at 1pm, actually, make it 2pm" produces "Let's meet at 2pm." Audio is sent directly from your Mac to the Gemini API using your own API key, with no intermediary server or account.
Is jot-gemini-transcribe-macOS free to use? What license does it have?
jot-gemini-transcribe-macOS is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is jot-gemini-transcribe-macOS still maintained?
Updated within 6 months. The last commit to jot-gemini-transcribe-macOS was on Aug 27, 2026.
What can I use jot-gemini-transcribe-macOS for?
Drafting emails, chat messages, or documents by speaking instead of typing. Dictating code-related names and technical terms with a custom dictionary. Capturing notes quickly on a Mac without switching windows.
How do I install jot-gemini-transcribe-macOS?
Download the latest Jot DMG from the Releases page Drag Jot into Applications and launch it from there Paste a Gemini API key from Google AI Studio and grant microphone and Accessibility permissions
Where can I get jot-gemini-transcribe-macOS?
The source code and installation instructions are on GitHub: https://github.com/google-gemini/jot-gemini-transcribe-macOS

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects