Skip to content
diktoappPublic

About

Voice-to-text for macOS. Records speech, transcribes locally. Whisper + Parakeet. No cloud.

Topics

Resources

Security policy

Stars

6 stars

Watchers

0 watching

Forks

Repository files navigation

Dikto

Dikto

Privacy-first voice-to-text for macOS. Records speech, transcribes locally. No cloud.

macOS 14+ (Sonoma) · Apple Silicon

Build Test Release Downloads License: MIT

Website · Download · Install

Website · Download · Install · Discussions

Quick Install

One-liner (downloads, verifies, installs, bypasses Gatekeeper):

curl -fsSL https://diktoapp.github.io/install.sh | bash

Homebrew:

brew tap diktoapp/dikto
# GUI app
brew install --cask dikto
# CLI only (builds from source)
brew install diktoapp/dikto/dikto

DMG: Download from Releases.

Gatekeeper: "Dikto can't be opened"

Dikto is ad-hoc signed but not Apple-notarized. The Homebrew and install-script methods bypass Gatekeeper automatically. If you downloaded the DMG directly:

  1. Right-click > Open: Right-click Dikto.app, choose "Open", click "Open" in the dialog.
  2. Terminal: xattr -cr /Applications/Dikto.app
  3. System Settings: Privacy & Security > scroll down > "Open Anyway".

Privacy

Your voice stays on your device. Dikto never connects to a server. All speech processing happens locally using on-device ML models. No accounts, no telemetry, no cloud APIs.

Usage

Press Option+Space to start recording. Speech is transcribed when you stop or silence is detected. The text is copied to your clipboard and pasted into the active app.

On first launch, macOS will prompt for Microphone access. Grant Accessibility permission in System Settings for auto-paste.

Models

Model Size Description
parakeet-tdt-0.6b-v2 (default) 2.5 GB NVIDIA Parakeet TDT — high accuracy English
parakeet-tdt-0.6b-v3 2.6 GB NVIDIA Parakeet TDT — 25 EU languages
whisper-tiny 75 MB Whisper Tiny — fast, 99 languages
whisper-small 460 MB Whisper Small — balanced accuracy & speed
whisper-large-v3-turbo 1.6 GB Whisper Large v3 Turbo — highest accuracy
distil-whisper-large-v3 1.5 GB Distil-Whisper — 6x faster Whisper

Download a model with the CLI:

dikto --setup --model whisper-small

Then select it in the app's Settings.

Architecture

  • Rust core (dikto-core) — audio capture, VAD, ASR engine, model management
  • Swift UI (DiktoApp) — SwiftUI menu-bar app with recording overlay
  • CLI (dikto-cli) — headless model setup

Config: ~/.config/dikto/config.json · Models: ~/.local/share/dikto/models/

Build from source

Prerequisites: Rust (1.75+), cmake (brew install cmake), macOS 14+.

git clone https://github.com/diktoapp/dikto.git
cd dikto
make build-app
open build/Dikto.app

License

MIT

About

Voice-to-text for macOS. Records speech, transcribes locally. Whisper + Parakeet. No cloud.

Topics

Resources

Security policy

Stars

6 stars

Watchers

0 watching

Forks

Releases

Packages

Used by

Contributors

Languages