For the complete documentation index, see llms.txt. This page is also available as Markdown.

Configure eSignet

Once an eSignet instance is up and running, tuning its behavior happens across a small, well-defined set of places — environment variables, a service YAML file, declarative UI assets, and per-client fields. All of it is documented directly in the eSignet repository, verified against the actual implementation, rather than duplicated here — configuration changes with the code, and a copy in this doc site would drift out of date. This page gives you the crux of each reference file and points you to the real one.

Where Configuration Lives

  • Environment variables and esignet-service/data/deployment.yaml — most service-level settings, resolved as env var > deployment.yaml > compiled-in default.

  • Declarative UI assets (<DATA_DIR>/{flows,layouts,themes,i18n}/*.yaml) — where ACR and login-ID behavior actually live, since neither is a deployment.yaml property.

  • Per-client rows in the client_detail table, set via the /client-mgmt/* API — claims, ACR values, redirect URIs, and other per-application settings.

  • oidc-ui's own runtime config (window._env_, theme/config.json) — build-time and runtime settings for the login/consent UI itself.

The Reference Files

Configuration: The Full Property Reference

The full list of environment variables and deployment.yaml settings for esignet-service and oidc-ui. Refer to it whenever you need an exact variable name, default value, or precedence rule.

See the full reference here (Link TBA) for every property and detail.

ACR: Which Authentication Methods a Client Can Request

Defines the fixed mosip:idp:acr:* values a client's authContextRefs can request, the ACR-to-AMR mapping advertised in discovery, and how each ACR value maps to an actual login screen (OTP, password, biometrics, KBI) in the flow YAML.

See the full reference here (Link TBA) for every property and detail.

Claims: What Personal Data a Client Can Receive

Defines the two-plane claim model — a per-client claims allow-list and a deployment-wide scope-to-claims mapping, with a claim released only if it appears in both — plus how ID token and userinfo delivery (signed vs. encrypted) is configured per client.

See the full reference here (Link TBA) for every property and detail.

Login ID: What Identifiers Users Can Log In With

Defines the shipped login-ID types — UIN/VID, mobile number, email, and NRC ID, plus a fifth KBI-only identifier — and each one's validation rule. Configured directly in the flow YAML, not deployment.yaml.

See the full reference here (Link TBA) for every property and detail.

Last updated

Was this helpful?