Product blog

Product blog

Product updates, field guides, and field notes—each with links to docs and the hub where relevant. High-level milestones also appear on the changelog.

  1. Why on-device AI is private by design

    Field guide · Jun 7, 2026

    Why on-device AI is private by design

    The difference between “we promise not to look” and “the data physically cannot leave.”

    • Structural, not policy
    • No upload, no log
    • The model comes to you

    NSS Background Remover

    Read
  2. What AI runs on your device versus in the cloud

    Field guide · Jun 7, 2026

    What runs on your device vs in the cloud

    A practical guide to telling on-device AI from cloud AI, and what each implies.

    • On-device: many tasks
    • Cloud: huge models
    • How to tell

    NSS Background Remover

    Read
  3. How AI runs in your browser with WebGPU and WebAssembly

    Field guide · Jun 7, 2026

    How AI runs in your browser (WebGPU and WebAssembly, explained)

    A plain-language explainer of the browser tech that makes on-device AI fast enough to use.

    • WebGPU uses your GPU
    • WASM is the fallback
    • No install needed

    NSS Background Remover

    Read
  4. The privacy cost of free AI tools

    Field guide · Jun 7, 2026

    The privacy cost of free AI tools

    What "free" can really mean when your images or audio are uploaded to a server.

    • Uploads = exposure
    • Read the terms
    • Free + private is possible

    NSS Background Remover

    Read
  5. Do browser AI tools work offline?

    Field guide · Jun 7, 2026

    Do browser AI tools work offline?

    How on-device AI tools keep working without a connection, and what still needs the network.

    • Works after first load
    • PWA + cached model
    • No connection needed

    NSS Background Remover

    Read
  6. An image and candidate labels scored by CLIP in the browser, with the best match highlighted

    Field guide · Jun 1, 2026

    What CLIP zero-shot tagging does on your device

    A plain explanation of CLIP zero-shot tagging, confidence scoring, and similarity search — and what it means that they run locally on your images.

    • Real CLIP, not keyword guessing
    • Zero-shot: no training step
    • Cosine similarity finds related images

    NSS Background Remover

    Read
  7. A user-hosted ONNX model URL loading into the browser and running with WebGPU and a WASM fallback

    Field guide · May 30, 2026

    Bring your own ONNX model: running your own weights in the browser

    The logical end state of on-device AI — if the computation happens on your machine, the model does not have to be ours. How BYO-ONNX works and why it exists.

    • Host your own ONNX at a URL
    • WebGPU primary, WASM fallback
    • Seven swappable capabilities

    NSS Background Remover

    Read
  8. Auditing browser AI tools: on-device processing versus quiet uploads to a server

    Field guide · May 27, 2026

    Is my data safe with browser AI tools? A plain-English audit

    A practical, non-technical guide to whether your data is safe with browser-based AI tools — the upload-versus-on-device distinction, what each model protects, and a checklist for auditing any tool.

    • On-device = files never uploaded
    • Check upload vs local
    • Free is not the same as private

    NSS Background Remover

    Read