Transcription always runs on-device. For summaries and the AI Assistant, you choose how the model runs: entirely on your device with a local LLM, with no data sent anywhere, or with your own API key from a cloud provider for higher-quality output.
Either way, your files and your chats stay on your device.



