Turso

Turso Database

SQLite-compatible embedded database for modern applications, AI agents, and edge computing.

Links

• Documentation
• Changelog
• GitHub

Quick Navigation

When to Use

• Embedded SQLite database with cloud sync
• AI agent state management and multi-agent coordination
• Offline-first applications
• Encrypted databases (AEGIS, AES-GCM)
• Edge computing and IoT devices

Core Concepts

libSQL

Turso is built on libSQL, an open-source fork of SQLite with:

• Native encryption (AEGIS-256, AES-GCM)
• Async I/O (Linux io_uring)
• Cloud sync capabilities

Deployment Options

1. Embedded — runs locally in your app
2. Turso Cloud — managed platform with branching, backups
3. Hybrid — local with cloud sync (push/pull)

Common Patterns

Encrypted Database

openssl rand -hex 32  # Generate key
tursodb --experimental-encryption "file:db.db?cipher=aegis256&hexkey=YOUR_KEY"

Cloud Sync

import { connect } from "@tursodatabase/sync";

const db = await connect({
  path: "./local.db",
  url: "turso://...", // also accepts "libsql://..." — both schemes work
  authToken: process.env.TURSO_AUTH_TOKEN,
});

await db.push(); // local → cloud
await db.pull(); // cloud → local

Agent Database

import { connect } from "@tursodatabase/database";

// Local-first
const db = await connect("agent.db");

// Or with sync
const db = await connect({
  path: "agent.db",
  url: "https://db.turso.io",
  authToken: "...",
  sync: "full",
});

Version

Based on product version: 0.7.2

Release Notes

0.7.1 – 0.7.2

• JS/serverless SDK: transactionAsync() is the new closure-safe transaction API. transaction() is now deprecated — its closure semantics are unsound once transactions can run concurrently, because statements captured over the outer db/conn instance can be scheduled out of order. Migrate transaction callbacks to use the txn handle passed into transactionAsync() instead of closing over the outer instance (see references/agents.md).
• Breaking: Connection.execute() was removed from the serverless driver's native-mirroring surface (connect()), which now matches @tursodatabase/database and exposes run/get/all/iterate/exec/batch/transaction(Async) instead. The libsql-compatible createClient() layer keeps its own execute() and is unaffected (see references/auth.md).
• Fixed a stale inTransaction flag after execute()/batch() that could leave a server-side write transaction open past a constraint error.
• Per-query requestHeaders, passed through the trailing query-options argument, let you attach custom headers (e.g. a request-identity header) to a single call instead of only at the connection level.
• Sync engine: the remote pull protocol (page-based WAL vs. MVCC logical-log) is now auto-detected on first contact, so the old logical_mvcc_pull flag becomes an optional manual override instead of a requirement. A WAL-mode local replica is automatically converted to MVCC journal mode in place when it syncs against an MVCC remote.
• Sync bindings (Rust, Python, JavaScript, Go, React Native) accept both turso:// and libsql:// remote URLs interchangeably.
• Core fixes: an IN (...) list query-cost regression that degraded InSeek to a full scan, a missing index left behind after UPSERT due to a pre-constraint-check index mutation, a change-count leak from sequences under MVCC, and DELETE/upsert replay bugs for tables with composite or non-rowid primary keys.

0.7.0

• SQL surface: SQL-standard scalar functions with PostgreSQL-compatible aliases, PostgreSQL-style sequences, MVCC-safe AUTOINCREMENT, window-function work (row_number() on VDBE aggregate machinery, FILTER in window clauses), and WITHIN GROUP ordered-set aggregates.
• MVCC/durability: passive checkpoint for MVCC, portable logical-log metadata for Turso sync, and Aristo WAL verification.
• Collations: core custom collation support and locale-backed collations.
• .NET / platforms: NativeAOT static linking, remote transactions and batches, a Turso EF Core SQLite provider, NuGet native targets, and Windows ARM64 CLI releases.

Earlier (0.6.0)

JS/serverless timeouts, interactive transactions, Python SQLAlchemy improvements (sqlalchemy-libsql), npm-based CLI distribution, and a broader SQL surface for local-first/agent workloads — still in effect, see references/agents.md and references/installation.md.