Skip to Content
DocsHow Valyd works

How Valyd works

New to Valyd or OIDC? Here’s the whole mental model in one picture: a person verifies once, the proof lives on their Valyd ID, and every app that adds Connect with Valyd just reads it.

The three pictures below unpack that loop.

1. A person verifies once

Somewhere — in your app, or any app using Valyd — a person proves who they are: they scan their ID, pass a liveness check, match their face, or have a professional license looked up. Valyd runs that check for real (there are no simulated results) and, when the person has a Valyd account, the passed outcome is saved to it as a proof.

2. They connect their identity to your app

Connect with Valyd is a standard OIDC button (its “Sign in with Valyd” form can double as your login button, like “Sign in with Google”). The person authenticates with their verified identity — face, not passwords — and approves what your app may read (the scopes). Your backend gets tokens.

You never see their documents — you read the answers: id_verified: true, verified licenses, age bands.

3. Need a new check? Attach it to the account

If the account doesn’t hold the proof you need yet, run the check yourself — with the signed-in user’s token attached, so the result saves back to their account:

That’s the loop: verify once → sign in anywhere → read the proof — and only re-verify when your policy wants a fresher answer.

The two products

Everything above is Reusable Verification: the person connects their Valyd identity via Connect with Valyd, your app reads the verified data it already holds, and runs a verification through a configured workflow for anything missing.

There is one deliberately separate product: the Unique Human API. It answers a single question — is this a live, unique human? — with nothing but your API key: Liveness stops a photo or replay standing in for a real person, and Uniqueness catches the same face opening a second account. No user login is involved, the result returns straight to your caller, and storing and protecting the person’s data is your responsibility.

What workflows let you build

The workflow checks aren’t just onboarding KYC — you run a verification whenever you need fresh proof, tied to the signed-in person:

  • Prove it’s really them, right now — a face match against their enrolled face before a sensitive action (a payout, a settings change, a shift clock-in).
  • Confirm they’re actually there — a location check proves the person is where they claim (home-visit care, field work, geofenced access).
  • Re-check a live credential — re-verify a professional license against the registry so an expired or revoked one is caught, not trusted from last year.
  • Confirm they’re a live human — liveness stops a photo or replay standing in for the real person.
  • Gate by age — an age band (is_18_plus, …) without ever touching their date of birth.

Compose several into one workflow and run them in a single session — each passed check saves as a reusable proof on the person’s Valyd ID, so next time you just read it. Verify once, then re-prove exactly what your policy needs, exactly when it needs it.

Where to next

Last updated on
OverviewCreate an account