Getting started

Introduction

Silker AI is a runtime security platform that inspects your application's traffic in real time - blocking attacks as they happen and streaming threat telemetry to your dashboard.

What is Silker AI?

Unlike scanners that analyze code or infrastructure at rest, Silker AI protects your application while it runs. It sits in the request path - as a Node SDK, a Next.js edge middleware, or a standalone reverse proxy - and inspects incoming requests and outgoing responses against a layered detection engine (OWASP Top 10, prompt injection, data-leak prevention, rate limiting, and more).

Detected threats are blocked at the edge and reported to the platform, where you can monitor activity, manage banned IPs, and tune detection from the dashboard without redeploying your app.

How it works

The flow is the same regardless of how you deploy:

Inspect - each request/response is scored by the detection engine in-process.
Block - malicious traffic is rejected with a 403 before it reaches your handlers (fail-open: if Silker errors, your app keeps serving).
Report - sanitized security events are batched and sent to the ingest endpoint.
Sync - the ingest response returns dashboard-managed config and the current banned-IP list, applied on the fly.

Silker AI does not currently expose a public REST API for developers to build against. The platform endpoints are internal to the dashboard and SDK. You integrate by installing the SDK or deploying the proxy - not by calling our API directly. See Data & privacy for the one contract the SDK uses.