Telemetry & Data Collection
Yeti Architecture Group ("Yeti," "we," "us") operates the headless widget infrastructure deployed on client environments. We log diagnostic telemetry, IP ranges, browser user agents, and conversation transcripts initiated through our JS modules.
Data is collected deterministically to fulfill the routing obligations of the service. We do not aggregate end-user data for shadow-profiling or external marketing purposes.
RAG Context Processing
When a client supplies documentation for Retrieval-Augmented Generation (RAG), that data is indexed into isolated vector databases. Yeti does not cross-contaminate RAG indexes between Client IDs. Your proprietary documentation remains cryptographically siloed and is never used to train generalized Large Language Models.
Data Retention & Purging
| Data Type | Retention Period |
|---|---|
| Conversation Transcripts | 30 Days (Rolling Purge) |
| Lead Payloads (Emails, etc.) | Stored until webhook delivery confirmation, max 72 hours. |
| System Error Logs | 14 Days |
Third-Party Routing
As an automation layer, Yeti acts primarily as a router. When you configure webhooks or integrations (e.g., sending leads to a CRM), you authorize Yeti to transmit payload data outside our secure perimeter. We are not liable for the data handling practices of the third-party endpoints you dictate.