Database API
库、表、字段、事件和行的 API 参考。
（待译）
Database is a headless content store: structured tables and rows that people, agents and scripts read and write directly through the API, the way a headless CMS works. Concepts and behavior follow Supabase (PostgREST filters, projections, ordering and row writes via supabase-js), so Supabase experience applies. Differences: storage uses SQLite, not PostgreSQL, and schema changes use table/field APIs rather than SQL migrations or raw SQL; access uses a Praxis credential scoped to its tenant, with no Supabase Auth or row-level security policies; realtime row subscriptions are unavailable; casts, aggregates, spread and join modifiers are unsupported, and filtered planned/estimated counts may be unknown; upsert targets id or one unique field rather than composite keys. For an application's own runtime data, use the application's own d1, kv or r2 components instead. Start with praxis.database.databases.create to create a database, then use its returned id for tables and rows.
用 API Key 作为 Bearer 凭据，通过以下地址管理库与表结构。
https://api.deepraxis.ai/database/v1