Skip to main content
Use a browser SDK to identify a visitor and obtain an Identification ID. Send that ID with the action your application is protecting. Your backend reads domain-scoped History for identification, risk signals and Risk Score, or verifies a signed webhook.
Only the Public Key belongs in browser code. Keep Private API Keys and webhook signing secrets on your server. Never put them in a public environment variable, browser bundle or GTM container.

Browser packages

See each package README for framework versions, provider setup, consent-aware loading and required peer dependencies. Follow the identification flow when passing an Identification ID to a protected server action. An Identification ID alone is not a risk verdict.

Server packages

The Node.js guide uses @shieldlabs-ai/node@1.3.0 for domain-scoped entity reads and bounded retrieval of a new identification. The other language guides combine the public HTTPS History API with the matching SDK’s raw-body webhook verification. Each guide pins the published package it tests. Resolve the domain’s public UUID with GET /api/v1/domains, then read lists or cards under /api/v1/domains/{domain_id}/history/. Preserve risk.state, signals.state, nulls and unknown fields. A successful lookup with unavailable risk is not a clean verdict. See History for filters and cursor pagination, and webhooks for signature handling.

OpenAPI and SDK updates

Download the OpenAPI YAML or JSON to inspect HTTP endpoints and generate types. Install the supported SDK package for your language separately, and review its changelog when upgrading. Lists support identifications and users. Cards support identifications, users, visitors, devices and IPs. Responses may also include related counts. Lists return an opaque next_page_token, without an exact total. A bounded page is not a complete domain-wide traffic report.