Aiko: AI-Powered Audio Transcription

Aiko, developed by Sindre Sorhus, is a high-quality on-device audio transcription tool available for macOS and iOS. It leverages OpenAI's Whisper model, running locally to ensure complete privacy by keeping audio data on your device. This makes it ideal for transcribing sensitive recordings such as meetings, lectures, and more.

Key Features

On-Device Transcription: Utilizes Whisper large v2 model on macOS and medium or small models on iOS, depending on memory availability.
Multi-Language Support: Handles audio in over 100 languages, catering to a wide range of users.
Export Options: Allows export of transcriptions as text, JSON, or subtitles.
Privacy-Focused: No internet connection is required, ensuring data never leaves your device.
Shortcuts Integration: Supports batch transcription and recording via iOS/macOS shortcuts for enhanced workflow.

Performance: Aiko prioritizes accuracy over speed, which may result in slower transcription times, especially on older devices. Performance is expected to improve with future updates, including potential CoreML support. It requires at least 16 GB of RAM on Macs and works best on Apple Silicon devices.

Usage: Users can record audio directly in the app or import compatible audio files via the share sheet on iOS or drag-and-drop on macOS. For advanced workflows, shortcuts enable automation, such as transcribing files directly from Finder on macOS.

Older Versions

1.8.2 for macOS 14
1.5.3 for macOS 13

These are free for everyone but they will not run on newer macOS versions.