Choose how to collect interactions from your websites, applications, and services. Start with the route that fits your infrastructure, then validate that the expected events reach Darwinium.
Recommended: integrate through your CDN
For traffic served through a supported CDN, CDN / edge integration is our preferred starting point. It lets you extend coverage to additional touch points through journey and route configuration, reducing the application changes needed to observe more of a journey.
Start with Edge architecture, then follow the guide for Cloudflare, AWS CloudFront, or Akamai.
Other integration paths
- Websites using tags: Install JavaScript tags, review JavaScript profiling behavior, and use the Event API for API-based event processing.
- Mobile applications: follow Mobile SDK Deployment and the Mobile SDK Reference. Mobile SDKs provide native app profiling; observing requests at the CDN does not replace that profiling.
- Backend services: set up Certificates & API access, then use the Event API.
These approaches can be combined. Your first successful event connects collection, journey configuration, deployment and verification.
Connect collection to a journey
Profiling collects device and behavioral data. A journey step defines the interaction to process, the attributes to map and the workflows to execute. Review Journeys, steps & events, then use the Journey Recorder or create a journey. Add profiling to the relevant pages or app flow and confirm how its data accompanies the protected request.
Validate the installation
Follow Deployment confirmation, then open Investigations to inspect the events from your test journey. If events or profiling are missing, use the installation troubleshooting guide.
Protect an HTTP MCP endpoint
Start with the CDN route above, then follow Monitor MCP and JSON-RPC traffic to map methods, tool names, arguments and results into journey events. Client-side browser profiling is not required to map an HTTP tool call. Validate your endpoint transport and response-inspection requirements against the deployment target.