For AI-agent builders

Let your AI agents act for Norwegian companies

Data APIs tell your agent about a Norwegian company. Apier lets your agent act for it — with delegated authority, an audit trail and Norwegian legal grounding, behind a single API. Ship the integration in days instead of months.

Why agents stall at the Norwegian border

Reading public registry data is easy. Acting on it — filing, verifying authority, submitting to a government endpoint — is where an autonomous agent runs into infrastructure it was never built to own.

• Your agent can't reach Altinn directly

  Altinn 3 is a machine-to-machine surface gated by Maskinporten. An agent can't just call it — it needs a virksomhetssertifikat, an RSA keypair, a Maskinporten client and an Altinn System Register entry before a single request authenticates.

• Delegation is harder than authentication

  Every company your agent acts for must grant a system-user delegation scoped to the right access packages. Those delegations expire, get revoked, and have to be re-granted on the organisation's side — state your agent has to track, not assume.

• Compliance is a moving target

  What a company must file, who may sign for it, and when it's due are defined by Norwegian law that changes. Hard-coding a deadline calendar or a signing-authority rule means shipping a bug the day the rule changes.

• Binding actions need an audit trail

  An autonomous agent that submits a filing needs a defensible record of what it did, under whose authority, against which rule version. Without that forensic chain, no compliance officer will let it act unsupervised.

One execution layer your agent can call

Apier sits between your agent and Norwegian government systems — Altinn, Brønnøysund, Maskinporten — and owns the parts that make acting safe: authentication, delegation, deterministic compliance, and an immutable audit trail. Your agent sends one API key; the rest is resolved in the execution layer.

• Authentication, handled

  Apier owns the Maskinporten token exchange and the System Register plumbing. Your agent sends one Bearer API key.

• Delegation, resolved

  Acting capacity — signaturrett, prokura, system-user scopes — resolved into a clear yes/no answer before any binding action is attempted.

• Compliance, deterministic

  Obligations and deadlines derived from a versioned Rulebook, computed in Oslo time. Same input + same rule version = same answer, every time.

• Audit, built in

  Every response carries trust _meta and lands an immutable provenance record, so what your agent did is reconstructable from a single correlation id.