Introduction to Keystone CDP
Keystone is a high-performance Customer Data Platform engineered specifically for handling aggressive data ingestion loops and unified cross-channel attribution.
Core Architectural Philosophies
Keystone was built to process extreme scale without schema lock-in.
- Schemaless Ingestion (JSONB): Data structures are volatile. We use PostgreSQL
JSONBalongside dynamic schema enforcers, not rigid columns. - Probabilistic vs Deterministic Resolution: Not all connections are certain. Keystone uses an Anchor/Shadow model for identity merging. Shadow profiles coalesce under Anchors deterministically when identities solidify.
- Micro-batch Optimization: Relying on arq/Redis queues, we process background enrichment without blocking API threads.
Security Overview
The entire API is secured via robust cryptographic headers:
{
"X-API-Key": "YOUR_SECRET_KEY"
}
Please refer to the API Reference for the complete endpoint structural map.