Download and switch models

Download models

Open Settings > Models.
Click Download next to a model that is not installed.
Wait for the progress bar to finish.

You can remove non-bundled models with Delete.

Choose your active model

In Settings > General, select a Local model size. If the model is not installed, you will see a prompt and SpeakLane will fall back to the closest available model.

English-only option

Enable English-only models to use .en variants that are tuned for English dictation.

Model tradeoffs

Tiny: fastest and smallest, good for quick drafts.
Base: balanced speed and accuracy.
Small: better accuracy for longer or noisy audio.
Medium: high accuracy, slower and heavier.
Large: best overall accuracy, needs more memory.
Turbo: very fast with near-large quality.

Tip

Larger models are more accurate but slower and require more memory. Try Tiny or Turbo for low latency.