Secure Local AI
A planned product category for local AI software. This page establishes disclosure requirements without presenting design considerations as implemented features.
Product disclosure
Not yet finalized
Exact behavior must be documented before any release.
- Type
- Local AI application
- Status
- PlannedNo public release or implementation is represented here.
- Version
- Not yet released
- Platform
- Not finalized
- Processing
- Not finalizedLocal inference is a design consideration, not a confirmed feature.
- Network behavior
- Not finalizedModel acquisition and optional or required runtime requests must be disclosed.
- Storage
- Not finalizedModel files, browser cache, and application storage are design considerations.
- Telemetry
- Not finalized
- License
- Not finalized
Design considerations
Questions the product must answer
A local AI application can have heavier runtime and storage requirements than a browser utility. The product must document these boundaries without being folded into the Web Utilities runtime.
- Whether inference runs locally, remotely, or in a hybrid arrangement.
- How models are acquired, verified, updated, and stored.
- Whether browser cache, WebGPU, native runtimes, or application storage are required.
- Which network requests are optional, required, or user-initiated.
- Whether telemetry exists and how users can understand or control it.
Source and documentation
Not publicly linked. No public Secure Local AI repository or release documentation is currently available through the organization catalog. Source, license, entry point, and technical documentation must be added when they are established.
Release disclosure will take precedence
These considerations are not promises about a future implementation. A product release must publish its actual processing, network, storage, model, telemetry, platform, and update behavior.