Turso

A SQLite-compatible database built on libSQL that syncs embedded local read replicas from a cloud primary for very low-latency reads.

Visit Turso →

Some links may be affiliate links — we may earn a commission if you sign up, at no extra cost to you. It never affects our rankings. Learn more.

Turso is a managed database platform built on SQLite that adds cloud features such as sync, embedded (local-first) replicas, and native vector search. It began as libSQL, an open-source C fork of SQLite that still powers Turso Cloud today, and is transitioning to "Turso Database," a from-scratch rewrite of SQLite in Rust (formerly codenamed Limbo). Its model treats a database as a lightweight file rather than a long-running server, enabling large numbers of small, isolated databases for per-user or per-tenant architectures.

Key features

SQLite-compatible SQL dialect, file format, and C API
Embedded replicas that keep a local read copy synced from a cloud primary for microsecond local reads
Native vector search for AI/RAG workloads (exact search today; approximate/ANN indexing still maturing)
Change-data-capture sync for local-first, offline-capable writes
Database-per-user model scaling to many isolated databases (unlimited on paid plans)
MVCC-based concurrent writes (BEGIN CONCURRENT) in the Rust engine

Specs

Category	Managed database
Engine	SQLite / libSQL
Pricing model	usage-based
Free tier	Yes
Regions	edge / multi-region
Self-hostable	Yes
Open source	Yes

Last reviewed 2026-07-16 — see our methodology. Specs, pricing and free tiers change often, so confirm current details on the provider before deciding.

Pros & cons

Our editorial assessment — not vendor-stated facts.

Pros

SQLite compatibility keeps the learning curve and migration effort low
Embedded replicas enable genuine offline/local-first apps with microsecond local reads
Lightweight database-per-tenant model scales cheaply to many isolated databases
Free tier with generous limits and no credit card required
The database engine — libSQL and the MIT-licensed Rust rewrite — is open source and self-hostable

Cons

The Rust rewrite (Turso Database) is still pre-1.0: SQLite compatibility is not yet complete and features like concurrent writes and vector indexing are still maturing
Usage-based billing on rows read/written can be hard to predict and a source of surprise costs at scale
SQLite lineage makes it a weaker fit for high write-concurrency or very large single datasets than Postgres or MySQL
Product direction has shifted in recent years (some earlier features, such as edge/geographic replicas, have been de-emphasized for new users) and the libSQL-to-Rust transition adds roadmap uncertainty; the ecosystem is also smaller than mainstream databases

Pricing

Tiered subscription plans (Free, Developer, Scaler, Pro, plus custom Enterprise) with usage-based overage charges across storage, rows read, rows written, and sync. The free tier requires no credit card. A recent change (mid-2026) lets paid plans create unlimited databases, billed by the number of monthly active databases rather than by total database count.

Who should use Turso?

Developers building SQLite-based, local-first, or multi-tenant (database-per-user) applications who want managed hosting with sync and built-in vector search. Less suited to workloads needing heavy concurrent writes or the full feature set of a traditional relational database like Postgres or MySQL.

Frequently asked questions

Does Turso have a free tier?: Yes.
How much does Turso cost?: Tiered subscription plans (Free, Developer, Scaler, Pro, plus custom Enterprise) with usage-based overage charges across storage, rows read, rows written, and sync. The free tier requires no credit card. A recent change (mid-2026) lets paid plans create unlimited databases, billed by the number of monthly active databases rather than by total database count.
Who should use Turso?: Developers building SQLite-based, local-first, or multi-tenant (database-per-user) applications who want managed hosting with sync and built-in vector search. Less suited to workloads needing heavy concurrent writes or the full feature set of a traditional relational database like Postgres or MySQL.