Developer Docs
Integrate Kadryn from code, choose the right ingestion path, and verify production traffic safely.
Kadryn Developer Docs help engineering and platform teams connect AI traffic to Kadryn safely.
Use this section to authenticate server-side workloads, route provider traffic through Gateway, send direct usage events, attach metadata, preserve traces, configure webhooks, and validate production readiness.
Choose your path
Gateway
Use Gateway when Kadryn must sit in the runtime request path and enforce policies before provider execution.
Start here: Gateway
Direct ingest
Use Direct ingest when you already call providers directly and only need to send normalized usage into Kadryn.
Start here: Direct ingest
Webhooks
Use Webhooks when your systems need to react to Kadryn events such as reports, alerts, exports or delivery changes.
Start here: Webhooks
Diagnostics
Use Diagnostics when you need to understand why setup, delivery or traffic visibility is not production-ready.
Start here: Diagnostics
Recommended implementation order
- Read the Overview.
- Configure Authentication.
- Create a Kadryn API key.
- Add or verify Provider keys if you use Gateway.
- Choose Gateway or Direct ingest.
- Add consistent Metadata.
- Preserve Tracing.
- Use Idempotency for retry-safe writes.
- Verify requests in Logs & Traces.
- Finish with the Production checklist.
Production principles
A production Kadryn integration should:
- keep Kadryn and provider secrets server-side;
- identify every workload with project, feature and environment metadata;
- make retryable writes idempotent;
- preserve request IDs, trace IDs and request group IDs;
- avoid logging secrets or sensitive payloads;
- verify traffic in Logs & Traces before broad rollout;
- use Diagnostics when readiness is degraded;
- document ownership, rotation and incident response steps.
Developer surfaces
API Keys
Create and rotate workspace API keys.
Open: API Keys
Gateway
Route provider traffic through Kadryn for runtime control.
Open: Gateway
Direct ingest
Send usage events without proxying provider calls.
Open: Direct ingest
Logs & Traces
Investigate requests, ingestion attempts, webhooks and diagnostics.
Open: Logs & Traces
Webhooks
Deliver Kadryn events to your systems.
Open: Webhooks
Diagnostics
Identify setup, traffic, provider, webhook and governance issues.
Open: Diagnostics
SDKs
Use starter examples and integration patterns.
Open: SDKs