#1297: WG New Spec: Verifiable Credentials Rendering Methods v1.0

Visit on Github

Opened Oct 29, 2026

Specification

https://w3c.github.io/vc-render-method/

Explainer

https://w3c.github.io/vc-render-method/#introduction

Links

The specification

Where and by whom is the work is being done?

  • GitHub repo: https://github.com/w3c/vc-render-method/
  • Primary contacts:
    • Dmitri Zagidulin (@dmitrizagidulin), Interop Alliance, Editor
    • Manu Sporny (@msporny), Digital Bazaar, Editor
    • Patrick St. Louis (@PatStLouis), Open Security and Identity, Editor
    • Hendry Poh (@pohhendry), Infocomm Media Development Authority of Singapore, Editor
    • Isaac Koh (@isaackps), Infocomm Media Development Authority of Singapore, Editor
    • Phil Archer (@philarcher), GS1, VCWG Chair
    • Brent Zundel (@brentzundel), Mesur.io, VCWG Chair
    • Ivan Herman (@iherman), W3C, Staff Contact
  • Organization/project driving the specification: VCWG Render Method Task Force; MIT Digital Credentials Consortium, Digital Bazaar, Infocomm Media Development Authority of Singapore, Government Technology Agency of Singapore
  • This work is being funded by: To be added
  • Primary standards group developing this feature: W3C VCWG
  • Incubation and standards groups that have discussed the design:

Feedback so far

You should also know that...

The specification is an extension to the Verifiable Credentials Data Model v2.1 via its reserved renderMethod extension point. It defines a single SuiteRenderMethod type that dispatches on a renderSuite value, with three render suites in scope for v1.0: static (an opaque, pre-rendered payload such as an image or an NFC binary payload), summary (a structured, localizable JSON summary the wallet renders natively), and html (an HTML, CSS, and JavaScript payload booted in a sandboxed iframe with the credential supplied as a data block). The html suite is the part most likely to interest the TAG: it runs issuer-supplied code inside the wallet and relies on the sandbox, a restricted set of capabilities, and host-mediated navigation to contain it. The threat model document was written with this in mind.

We are planning to enter Candidate Recommendation as soon as we have received horizontal reviews and responded to requests for changes. This is one of the reviews that is gating our entrance into Candidate Recommendation. Our goal is to enter Candidate Recommendation in Q1 2027.

<!-- Content below this is maintained by @w3c-tag-bot -->

Track conversations at https://tag-github-bot.w3.org/gh/w3ctag/design-reviews/1297

Discussions