Identity and zero trust
ConceptProductSince version 0.3.0
Product guide: what the capability brings, for whom, its limits and the version that delivered it - no configuration syntax.
| What it does | Verifies each caller's identity against your identity provider, and requires exactly what the contract asks, operation by operation. Four first-class identities: an OIDC token (a JWT or an introspected opaque token), an API key, a client certificate (mTLS), the headers of a trusted mesh. Closed by default: an undeclared header does not reach the upstream, an undeclared parameter is refused, so is a browser origin not allowed |
| Why it matters | Security no longer depends on each API team. The caller's tenant comes from a verified source (a signed claim, a key, a certificate), never from a header the client writes itself. Every authorization decision can be written to an audit log, subject hashed |
| Use cases | One provider for employees and another for customers, in the same gateway; API keys for a partner or a batch, bound to a tenant, operations and an end date; a bank's server presenting a certificate; renaming an audience without cutting the old tokens |
| Limits | Flexway consumes tokens, never issues them; sender-constrained tokens (DPoP, mTLS RFC 8705) are not built; verifying incoming webhook signatures and IP address lists come in 1.8.0 |
| Since | 0.3.0 (OIDC, closed by default), 0.1.0 (API keys, client certificates), 1.0.0 (opaque tokens, decision audit, several audiences) |
| Configure | Wiring your identity provider, API keys, Client certificates (mTLS) |