The marketing/landing site for CredentAgent — the consent layer for AI agents.
An AI agent proves a verifiable credential from the user's wallet before a consequential action completes. Identity leads; payments is one application. Library: https://github.com/openmobilehub/credentagent
A single, self-contained static page (index.html) — inline CSS + vanilla JS, zero runtime
dependencies. No framework, no bundler, no build step. It renders identically opened directly
from file://.
Open index.html in a browser. That's it. (Or python3 -m http.server and visit the page.)
GitHub Pages, via .github/workflows/pages.yml, on every push to main. Enable Pages once in
Settings → Pages → Source: GitHub Actions.
The "Honest by design" section mirrors the SDK's trust model. It must never claim a stronger
guarantee than the library actually provides. The canonical source is
docs/reference/trust-model.md in https://github.com/openmobilehub/credentagent. When the SDK's
trust_level advances (e.g. to issuer-verified), update the trust table here to match — the site
follows the SDK, never the reverse. Presence-only rails are labeled presence-only and are never
presented as a real safety control.
Apache-2.0 · An Open Mobile Hub project (Linux Foundation / OpenWallet Foundation).