# Substance — `architecture/database` **Output path:** `{kbFolder}/architecture/database.md` **Topic focus:** the persistence layer — storage engine(s), entities/tables, their relationships, field types, indexes, and migration mechanism. Describe *how data is stored at the technical level*. **Discovery input to read:** the `database` domain findings from Phase 1 (schema files, ORM models, migrations, connection config). **Required output:** - Confidence header on line 1. - Storage engine(s) and connection/config location. - Entities/tables with field types and key relationships. - Migration mechanism and where migrations live. - Technical field types — conceptual business meaning belongs to `domain-model`. **Not applicable:** if the project has no persistence layer, the topic is *absent* — a certainty, not a low-confidence guess. Write a **confident** not-applicable stub: line 1 is ``, then the `# Database` heading and the one-sentence absence statement `No persistence layer — this project stores no data; it is stateless.` Do NOT lower confidence and do not invent tables — an absent topic is known with certainty.