Documentation · v0.1.0
Zanith documentation.
A SQL data engine that doesn't generate code. Schema parses once at startup into a runtime graph; queries compile to parameterized SQL on the way out. These docs cover installation, schema, queries, advanced features, and the full reference.
Getting Started
Install the engine and run your first query in under a minute.
Schema
How models, fields, relations, and enums get into the runtime graph.
- →OverviewWhat the schema graph holds, how it's built, when it runs.
- →Models & fieldsScalar types, modifiers, field-level attributes.
- →Relations1:1, 1:N, M:N — all expressed in the schema graph.
- →EnumsLiteral value sets, default mapping, validation.
- →.zanith DSLThe file-based schema syntax, parser semantics.
- →MixinsReusable field bundles — id, timestamps, audit.
Queries
The model API: CRUD, filters, relations, aggregates.
- →OverviewThe shape of a typed query, how it compiles.
- →CRUDfindMany / findFirst / findUnique / create / update / delete.
- →Filters & whereEvery operator, with shape and compiled SQL.
- →Relationsinclude, nested where, multi-hop joins.
- →AggregatesgroupBy, count, sum, avg, having.
- →Insert & upsertRETURNING semantics, ON CONFLICT, bulk inserts.
Advanced
Transactions, raw SQL, JSONB, window functions, subqueries.
Studio
The bundled web UI. Workspace + database, eight tabs, three deployment modes.
- →OverviewnewTwo levels, eight surfaces, in-process with the engine.
- →Workspace levelnewCross-database — connections, activity, search.
- →Tables surfacenewBrowse, edit, bulk operations, sensitive masking.
- →SQL playgroundnewMulti-tab editor, plan history, watch, charts.
- →Security tabnewLock graph, RLS, roles, permission matrix.
- →CatalognewViews, sequences, triggers, functions, indexes.
- →Launch + authnewThree deployment shapes — local, staff, viewer.
Migrate
Generate, plan, verify, apply — with risk scoring and a recovery layer.
- →OverviewnewThe system in summary, with the contrast vs other tools.
- →Lifecyclenewgenerate · plan · verify · apply · audit, with failure paths.
- →Risk modelnewSix levels, twenty reason codes, three CI gates.
- →Shadow-DB verifynewThe gate that catches drift before up touches production.
- →Recovery (Phase 2)newSoft-drop, archive, restore, export, purge.
- →Audit + historynewFour bookkeeping tables, the history CLI, snapshot replay.
Reference
Configuration, type system, command-line tooling.
v0.1.0 · early
Some pages on this surface are stubs for features that are on the roadmap, not yet shipped — they're marked planned in the sidebar. Everything else describes the engine as it ships today. The full status list lives on /roadmap.