Turso's fork adding server mode and replication. SQLite accepts no external PRs.
SQLite is an embedded relational database engine that requires no server process and stores the entire database in a single file. It's the most widely deployed database in the world, used in smartphones, browsers, operating systems, and countless applications. libSQL extends it with server mode, replication, and edge computing capabilities while maintaining full API compatibility.
SQLite is one of the most deployed pieces of software in human history — it runs on literally billions of devices. It's also one of the most unusual open-source projects: the code is in the public domain, but the project accepts zero external contributions. Not "we're picky about contributions." Zero. Dr. D. Richard Hipp, SQLite's creator and benevolent dictator, has maintained this policy for over two decades, running the project through his company Hwaci with a small team of developers.
Glauber Costa, founder and CEO of Turso, saw this as both an incredible foundation and a frustrating limitation. In October 2022, he published the libSQL manifesto, arguing that SQLite needed to evolve for modern cloud and edge computing use cases — distributed databases, asynchronous I/O via io_uring, eBPF optimization, user-defined functions in WebAssembly — but that SQLite's closed contribution model made this evolution impossible.
libSQL launched with no code changes: just a manifesto, a Code of Conduct, an MIT license, and a list of ambitious directions. The audacity of forking one of the most successful and stable pieces of software ever written raised eyebrows. But Costa's team backed it up with substance, adding server mode, replication capabilities, and the features that SQLite's strict scope had always excluded.
The project maintained a conciliatory tone throughout, explicitly stating that "if and when SQLite changes its policy to accept contributions, we will gladly merge our work back into the core product." This wasn't a hostile fork — it was more like a parallel universe where SQLite had an open contribution policy.
libSQL has since become the foundation of Turso's database-as-a-service platform, gaining significant traction in the edge computing space. It demonstrates that even the most successful open-source projects can be limited by governance choices, and that "open source, not open contribution" leaves a gap that others will fill.
D. Richard Hipp creates SQLite with a no-external-contributions policy
Glauber Costa publishes the libSQL manifesto, forking SQLite
Initial fork released with no code changes — just manifesto and governance
libSQL adds server mode and replication — features SQLite's scope excludes
Turso's platform built on libSQL gains significant traction in edge computing
libSQL established as the standard for distributed SQLite use cases
libSQL carved out a new category: distributed, server-capable SQLite. By adding the features that Hipp's strict scope excluded — replication, server mode, extensibility — libSQL made SQLite viable for use cases its creator never intended to support. Turso's commercial success on top of libSQL validated the approach.
More philosophically, libSQL challenged the assumption that a project's creator always knows best about its scope. SQLite's "open source, not open contribution" model works brilliantly for an embedded database, but it also created an opportunity for a fork that embraces community input and modern architectures.