01 Can stay on device
Local speech recognition
Configure a local speech model and AI dictation can process recording and transcription on your device, even when the network is unavailable.
Privacy
Private communication, on your terms.
WiseMe is a local-first communication agent. Speech can remain on your device, model calls can go directly to the provider you choose, and your knowledge, memory, dictionary, and settings stay local.
Built for AI dictation and AI Draft without turning your everyday communication into data stored in WiseMe's cloud.
Could you make this reply warmer and clearer?
AI Draft reads screen context only when you trigger it
AI Draft
Here is a clearer version, keeping your tone and intent intact.
Can stay on device Local speech recognition
Direct connection Bring your own provider
Model & data routes
You choose where each step is processed.
WiseMe keeps each route clear: process speech on your device, connect your own AI provider, or choose a WiseMe cloud model.
02 Direct connection
Bring your own provider
Use your own API key when you want requests to go from your device directly to the model provider you choose. WiseMe does not store that content on its servers.
03 No retention
WiseMe cloud model
When you use WiseMe's built-in cloud model, only the content needed for the result is forwarded to a model provider. WiseMe does not store it or use it to train models.
WiseMe runs locally on your computer.
The core app runs on your device. Your knowledge base, personal memory, dictionary, and settings stay local, so WiseMe does not build a cloud profile from the everyday context that helps you communicate.
Knowledge base and personal memory stay on your device.
Custom dictionary and configuration stay on your device.
Voice, text, and screen content are not used to train WiseMe models.
Privacy controls
Private by default. Explicit when it matters.
WiseMe reads sensitive context only in response to an explicit action from you. The product makes that moment clear instead of hiding it in the background.
Microphone
Used only while AI dictation is active.
Screen context
Read only when you trigger AI Draft.
Diagnostics
Optional, anonymous, and never includes message content.
Training
Your content is not used to train WiseMe models.
Common privacy questions
Short answers for how voice, screen context, model calls, and memory are handled.
contact@wiseme.app
Can AI dictation work completely locally?
Yes. With a local speech model configured, recording and transcription happen on your device and can continue to work offline.
When is screen context read?
Only when you trigger AI Draft. It is used to understand the current screen and draft a response, then follows the model route you selected.
What happens with my own provider key?
Requests go from your device directly to the provider you choose. WiseMe does not store that content on WiseMe servers or use it for training.
What happens with the WiseMe cloud model?
The content needed for generation is forwarded to a model provider to complete the request. WiseMe does not store the content after the action or use it to train models.