Turso Builds a Postgres-Compatible Database Engine on Its Rust Rewrite of SQLite
Turso, which rewrote SQLite from scratch in Rust, is now building a Postgres-speaking frontend on the same engine, aiming to be what its CEO calls 'the LLVM of databases.'
Overview
Turso, the database company behind a full rewrite of SQLite in Rust, says it is now building a version of Postgres on top of that same engine. In a blog post co-authored by CEO Glauber Costa and co-founder Pekka Enberg, the company wrote: “We will write a modern version of Postgres, in Turso (which is itself written in Rust).”
What We Know
Turso’s core database engine began as what The Register describes as a project “initially codenamed Limbo” — a full, from-scratch rewrite of SQLite in Rust. According to the outlet, Costa and Enberg previously worked as engineers at ScyllaDB before starting a company called ChiselStrike in San Francisco, which forked SQLite as libSQL; “ChiselStrike had already pivoted to a SQLite-based cloud service and renamed itself Turso,” The Register reported.
In its own blog post, Turso describes the resulting engine as “or was until recently, exclusively a full rewrite of SQLite in Rust using a modern architecture.” The company says that engine already goes beyond stock SQLite in several respects: “It supports concurrent writes using MVCC (like Postgres), it has a rich type system (like Postgres), it has support for Materialized Views (unlike Postgres, those views actually auto-update!)”
The architectural link to SQLite runs deep. Turso explains that SQLite “has a very unique design: it compiles SQL (in the SQLite dialect) to its own bytecode language, called the VDBE.” The Register confirms Turso carried that same approach into its rewrite: “They built Turso using the same virtual machine design, one based on VDBE.” Costa framed the appeal of that design in blunt terms, telling The Register: “Look close enough, and every SQL database is just a fancy collection of B-Trees with a bunch of Indexes.”
That bytecode-VM architecture is what Turso says makes a Postgres frontend feasible on the same core. Both the company’s own post and The Register quote an identical line from Costa laying out the ambition: “Turso is becoming the LLVM of databases. One modern, reliable core; many database frontends compiled down onto it.” The Register reports that “this month, Turso-the-company launched the Postgres effort,” placing the start of the work in July 2026.
What We Don’t Know
Turso is explicit that it is not chasing a drop-in replacement for Postgres. The company writes: “Unlike SQLite, our current belief is that we need to be compatible enough, especially at the core functionality, but not really 100%.” The Register puts it similarly, reporting that Turso’s planned database “is not promising complete compatibility with the OG Postgres.” Neither the blog post nor The Register’s reporting specifies a release date, a compatibility benchmark, or which Postgres features will be supported first. The Register’s article does not disclose any funding figures or investor names tied to the Postgres effort.
Analysis
Turso’s framing casts the company less as a database vendor and more as a runtime-compiler project: a single bytecode core that different SQL dialects, SQLite’s and now Postgres’s, get compiled down onto. Costa’s own gloss on that philosophy leans toward patience over speed: “Slow-and-correct is exactly why we will still be here, and still be right, in three years,” he wrote on the company blog. Whether a Postgres-speaking layer built on a VDBE-style virtual machine can match the wire-level behavior developers expect from Postgres extensions and tooling remains, by the company’s own account, an open question it has only just begun answering.