OSS Tanbou
View on GitHub

Jarvis AI Assistant — Local-first macOS voice dictation that types into whatever app is active

OSS health 79
About these scores

Discovery score is an unbounded weighted, log-compressed index of stars, watchers, forks, and contributors. Growth momentum is its change over the observed period; OSS health is a 0–100 score from available repository recency, Community Health, and release data.

Stars
634
Primary language
TypeScript
License
MIT
Repository last updated
Aug 21, 2026

Overview

Jarvis AI Assistant is a macOS voice dictation and assistance layer: hold Fn, speak, and cleaned text is typed into the active application. Built with Electron, TypeScript, and React, it can use local Whisper/SenseVoice/Sherpa-ONNX transcription and Ollama, or cloud paths such as Deepgram and Gemini. The latest stable release checked is v1.3.11.

Features and best fit

Based on official documentation; not hands-on tested · Content checked:

Connect a global Fn shortcut to recording, transcription, cleanup, and typing

Holding Fn records speech, then filler removal, punctuation, grammar correction, or rewriting can be applied before text is inserted into the active application. The Electron app uses native addons for microphone, global-key, and typing integration.

Sources: [2][3]

Switch between local transcription/LLMs and cloud services

The README documents local Whisper, Parakeet, and Ollama paths. Release v1.3.11 adds offline SenseVoice Small through Sherpa-ONNX/CoreML, while Deepgram plus Gemini remains an option when cloud speed or accuracy is preferred.

Sources: [2][4][3]

For Mac users who want to replace everyday typing with voice, not just transcribe meetings

Jarvis is useful when the same dictation shortcut should work in editors, browsers, mail, and other apps. Local models can keep audio and text processing on-device for users who want to avoid subscription-based cloud dictation.

Sources: [2]

Review macOS permissions, local-model resource cost, and the gap between stable releases and main

The app needs powerful OS permissions such as microphone and Accessibility, with screen capture used by some capabilities. Local AI post-processing can add noticeable latency depending on hardware. The latest stable release is v1.3.11 from June 2026 while main has moved forward, so README features may be newer than the stable binary. Windows remains roadmap work.

Sources: [2][4][3]

Official sources

  1. [1]akshayaggarwal99/jarvis-ai-assistant — GitHub repository metadata(2026-09-17)
  2. [2]Jarvis AI Assistant — README(2026-09-17)
  3. [3]Jarvis AI Assistant — package.json(2026-09-17)
  4. [4]Jarvis v1.3.11 release(2026-09-17)
Supplemental curator note

Treat Jarvis primarily as a local-first dictation and voice-assistance layer that types into other apps, rather than as a general autonomous agent.

Try it in 3 steps

  1. 1

    Get the source

    git clone --depth 1 https://github.com/akshayaggarwal99/jarvis-ai-assistant.git
  2. 2

    Enter the repository

    cd jarvis-ai-assistant
  3. 3

    Check the official steps

    Continue with the commands in the README Installation, Quick Start, or Getting Started section.

    find . -maxdepth 1 -iname 'README*' -exec sed -n '1,220p' {} \; -quit
Check the official README

Growth

Growth trends · Last 30 days

634 Stars

Trend data is still being collected.

Development activity

Last 90 days · weekly

Commits (last 30 days)
1
Open PRs
1

Development activity is still being collected.

Built with

Categories and tags

Categories

Tags

GitHub data

GitHub dataView detailed GitHub data
View on GitHub

GitHub Topics

  • ai
  • assistant
  • dictation
  • jarvis
  • mac
  • voice
  • ironman
  • ironman-mode
  • jarvis-ai
  • jarvisai
Stars
634
Forks
108
Watchers
8
Open issues
5
Primary language
TypeScript
License
MIT
Repository last updated
Aug 21, 2026
Write a related article

Share a guide or use case for this OSS in Markdown. Articles are published after administrator approval.

Report incorrect information

Tell us if any listing information is incorrect or outdated.