Concepts

Radicas is easier to use — and much easier to build against — once you know how it models data. Four pages cover the whole mental model; each stands alone, but they read best in order.

The four pages

• The three layers — storage engine, cost canonical schema, and operational telemetry standard are three separate decisions, never conflated. Everything else builds on this split.
• Estimated vs authoritative cost — two standards in one engine: real-time estimated cost from telemetry, authoritative cost from invoices, reconciled span by span — plus the allocated basis that never reconciles.
• Tenancy and ingestion — one ingestion path for every source, with tenant identity resolved from your API key at the edge and impossible to spoof from the client.
• Features, users, and allocation — how radicas.feature and user context drive per-feature profitability, and how rate cards and share-of-seat allocation make seat spend attributable.