Turso

Turso is an embedded database engine that runs anywhere — on servers, in browsers, or on-device. It's a drop-in replacement for SQLite, rewritten in Rust for concurrent access and async I/O.

This package provides the turso CLI — an interactive SQL shell, a local sync server, and an MCP server for AI assistants.

Install

npm install -g turso

Or run directly without installing:

npx turso

Quick Start

# Start an interactive shell with an in-memory database
npx turso

# Open or create a database file
npx turso myapp.db

# Execute a SQL statement directly
npx turso myapp.db "SELECT * FROM users;"

Features Beyond SQLite

Turso is a drop-in replacement for SQLite, but adds features that SQLite doesn't have:

• Concurrent Writers — BEGIN CONCURRENT allows multiple writers without blocking, powered by MVCC
• Native Vector Search — vector32/vector64 types with distance functions (vector_distance_cos, vector_distance_l2)
• Change Data Capture — track row-level changes per connection with PRAGMA capture_data_changes_conn
• MCP Server — run as a Model Context Protocol server for AI assistants (--mcp)
• Local Sync Server — serve a database over HTTP for client SDKs to sync against (--sync-server)
• Array Types — array columns in STRICT tables with operators like @>, <@, ||
• Built-in Extensions — crypto, regexp, fuzzy matching, IP address functions, CSV, percentile

Experimental Features

These features are available behind --experimental-* flags:

• Materialized Views — incrementally maintained views with automatic change tracking
• Custom Types — user-defined types with CREATE TYPE, custom encode/decode and operators
• At-Rest Encryption — transparent database encryption (AES-GCM, AEGIS ciphers)
• Full-Text Search — Tantivy-powered FTS with custom index methods
• Generated Columns — virtual and stored computed columns
• Triggers — CREATE TRIGGER / DROP TRIGGER
• Attach — ATTACH DATABASE / DETACH DATABASE
• Autovacuum — automatic database compaction

Run npx turso --help for the full list of flags.

Examples

Interactive Shell

npx turso myapp.db

turso> CREATE TABLE users (id INTEGER PRIMARY KEY, name TEXT, email TEXT);
turso> INSERT INTO users VALUES (1, 'Alice', '[email protected]');
turso> SELECT * FROM users;
┌────┬───────┬───────────────────┐
│ id │ name  │ email             │
├────┼───────┼───────────────────┤
│  1 │ Alice │ [email protected] │
└────┴───────┴───────────────────┘

One-Shot Queries

# Run a query and exit
npx turso myapp.db "SELECT count(*) FROM users;"

# Pipe-friendly list output
npx turso -q -m list myapp.db "SELECT * FROM users;"

Embedded Database

Use Turso directly as an embedded database in your Node.js application with @tursodatabase/database:

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

const db = await connect("local.db");

await db.exec(`
  CREATE TABLE IF NOT EXISTS users (id INTEGER PRIMARY KEY, name TEXT, email TEXT);
`);

const insert = db.prepare("INSERT INTO users (name, email) VALUES (?, ?)");
await insert.run(["Alice", "[email protected]"]);

const select = db.prepare("SELECT * FROM users");
console.log(await select.all());

Local Sync Server

Start a local HTTP server that implements the Turso sync protocol. The @tursodatabase/sync SDK can sync against it:

npx turso myapp.db --sync-server "0.0.0.0:8080"

MCP Server

Start an MCP server so AI assistants can query your databases:

npx turso --mcp

Shell Commands

Inside the interactive shell, use .commands for database operations:

Supported Platforms

Links

• GitHub
• Documentation
• Discord

License

MIT