Garmingo Voice

Garmingo Voice is a private desktop application that provides real-time microphone enhancement, system-wide dictation, and meeting recording with searchable transcripts — described on the product page as "three flows, one app" for voice enhancement, dictation, and meeting transcripts.

Garmingo Voice is available for macOS and Windows; the product page states macOS & Windows are fully available and refers users to the system requirements table for supported versions.

The product page lists supported OS versions as Windows 10/11 or macOS 11 (Big Sur) and later. It shows minimum vs recommended resources (example values on the page): RAM minimum 500 MB / recommended 8 GB; disk minimum 500 MB / recommended 10 GB; VRAM none / recommended 10 GB.

Garmingo Voice is on-device by default and is described as able to run locally so features can work without a network connection. The product page states on-device dictation and transcription continue to work offline for several days before a quick recheck, and there is an optional cloud mode you can opt into for additional speed.

The product page states Garmingo Voice supports 99+ languages for both dictation and meeting transcripts, with automatic detection of language and accent.

According to Garmingo's privacy policy, local processing uses models on your device so audio and transcription data are processed exclusively on your machine and not transmitted. If you use cloud-based AI features, audio is transmitted to the cloud AI provider (identified in the privacy policy as Mistral AI) under GDPR-compliant terms. Garmingo also describes model-training participation as an explicit opt-in program; training on user data is not performed by default and any contributions for training are voluntary and require express consent.