Language model processing for the assistant's reasoning and drafting.
The AI model providers alfred_ relies on, and how we protect your data.
alfred_ uses third-party AI providers to deliver its features. They're set up so your inputs and outputs aren't used to train their models.
Last updated: November 3, 2026
How we handle AI provider data use
alfred_ reaches AI providers through their APIs. Each is set up so customer inputs and outputs aren't used to train third-party models.
No training on your data
Provider terms bar them from using customer inputs or outputs for model training.
Conditional providers
Some providers are used only for specific features. Each card below says when.
Verified settings
We check each provider's no-training status through its terms, settings, or contract.
AI model providers
These AI model providers may process customer data when delivering alfred_'s features.
Search, text-to-speech, and voice. Used when those features need it.
Background processing on open-weight models, run in alfred_'s own AWS account in the US.
Language model processing for email triage and briefings.
Routes the decision model used in text messages and the web app, with zero data retention.
Routing and conditional providers
Some requests may use fallback or feature-specific routes. Where a routing layer is used, it sends customer data only to providers whose terms match our no-training commitment.
Acceptable use
You may not use alfred_ or its AI features for illegal activity, security abuse, privacy violations, deceptive impersonation, or harmful content.
Scope
This page covers AI model providers. Other vendors are on our subprocessors page. Apps you choose to connect are governed by your relationship with those services and their own terms.
Questions about AI providers or data use?
If your organization needs more detail on a provider or how we handle data, we're happy to help.
Contact us