Documentation menu
Start here

Build with Qirava

A practical guide to the database, its query language and the contracts that connect your application to authenticated storage.

On this page

Start with the working core

Qirava provides a Rust embedded library and a standalone database server. Both enter the same transaction engine for records, QQL, indexes, object references, durable receipts and optional at-rest encryption. This documentation describes the development implementation.

Create your first database →

Choose a client driver →

Learn the model

  • A collection holds keyed JSON documents. A stored record includes its key, opaque revision and value.
  • One execute request is an ordered atomic transaction. Later operations can use earlier staged changes.
  • Declare secondary indexes explicitly. Use query bindings for values and retain exact request IDs for uncertain writes.
  • Object uploads can publish an immutable manifest together with a document reference.

Transactions and retries →

QQL reference →

Know the current boundary

Cluster consensus, read replicas, sharding, online bidirectional migration, full custody ceremonies and custom role administration remain production delivery work. Current capability reporting must not be read as a promise of those features.

Operating modes and availability →

Read the measurements →