Registry Evidence overview

If another service needs to check something your institution’s records settle, such as whether a person is an adult, Registry Evidence lets you answer that one question without sharing the record.

What a caller gets

The caller names a fixed question by its Requirement URI, such as urn:example:requirement:adult-status:v1, and the subject it is about. It gets back an assertion: a signed answer it can verify offline, carrying only the values the question declares. Asked whether a person is an adult, Registry Evidence reads a birth date from your system and returns true; the date stays with you.

What you define and run

You write an Evidence project, kept in version control: the questions, the fixed reads of your own systems that answer them, and synthetic cases that prove each answer offline. evidencectl checks and packages it. evidence serves it, signs each answer, records who asked in an audit trail, and refuses any caller, purpose, or value the project does not declare. How Evidence works explains each part and each refusal.

When it is not the right product

If the other service needs the record itself, or needs to change it, hold the record in Base Registry Engine and give that service an access profile there. If you need credentials with a lifecycle, such as revocation, look elsewhere: an SD-JWT VC answer from Registry Evidence is one more encoding of the same signed answer, with no lifecycle behind it.

Where to start

Your first assertion installs the tools, runs the example project over a synthetic source, and verifies one signed answer in about fifteen minutes. How Evidence works gives the model behind it.

Next

• Your first assertion