Adapters & databases

An adapter connects @valv/core to a specific database. Everything above the adapter (the query grammar, validation, policy injection, and the tool layer) is database-agnostic and lives in core. This page lists the supported databases and explains what an adapter does.

Supported databases

Prisma

The Prisma adapter reads its schema from your generated Prisma client, so you don’t introspect or hand-define a catalog. Create valv with the client:

import { createValv } from "@valv/prisma"

const valv = await createValv(prisma, { defaultPolicy: "deny-all" })

Prisma databases support all three write operations: create, update, and delete.

ClickHouse

The ClickHouse adapter connects with a @clickhouse/client instance. Pass a database name and either introspect the live schema or hand-define one:

import { createValv } from "@valv/clickhouse"

const valv = await createValv(client, {
  schema: "introspect",
  database: "analytics",
  defaultPolicy: "deny-all",
})

ClickHouse adds dialect functions to the query grammar, such as toStartOfInterval and quantileTiming. For writes, it supports create (insert) only.

What an adapter does

The SQL emitter is shared, so an adapter doesn’t write a compiler. It provides three things:

• introspect() produces a catalog from the live database.
• A small Dialect describes how to quote identifiers and bind parameters.
• execute() runs the compiled SQL and returns rows.

Because core owns validation, policy, and emit, every adapter inherits the same safety guarantees.

Next steps

• Schema: introspect or hand-define a catalog.
• Queries: the grammar shared across every adapter.