Local AI dictation, on your own device.
Verenu can transcribe and clean up dictation entirely on-device, with no cloud provider involved. Pick a local transcription model and a local cleanup model (or turn cleanup off), and dictation stays on your machine after the model files download.
Going fully on-device
Pick a local transcription model
Settings -> Models lists the on-device speech-to-text models. Download the one you want to use.
Pick a local cleanup model, or turn cleanup off
A downloaded local LLM cleans up filler words and formatting on-device. Skip this step and Verenu keeps the raw transcript as-is.
Dictate as usual
Hold the hotkey and speak. With both steps local, the dictation stays on your device after the models finish downloading.
Mixing a local transcription model with a cloud cleanup provider keeps audio on-device but sends the transcript to that provider. Full walkthrough in the local transcription docs.
Platform limits
Local model downloads and inference are available on Windows and Apple Silicon Macs. Intel Macs are gated off for now, since local model inference hasn't been validated on that hardware yet. Linux isn't a supported desktop target.
Speed and output quality depend on the model you pick and the machine running it. Larger local cleanup models need more memory and take longer to respond than the cloud providers listed on the BYOK page.