 | Newsletter · October 2026 |
| | Turso is joining SupabaseToday we’re sharing big news: Turso is being acquired by Supabase. Together, we’re building the database platform for the agentic era: everything agents need, from lightweight databases to Postgres at petabyte scale. What this means for you: nothing changes in how you use Turso. Your databases, APIs, and workflows keep running as they do today. Turso Database stays open source and actively developed. And when SQLite isn’t enough, you’ll have a clear path into standard Postgres through the Supabase ecosystem. The team is staying too. Glauber joins Supabase as Head of Agentic Services, alongside Pekka and the rest of the Turso team. Expect deeper integration between Turso and Supabase in the coming months. |
| // What’s new | | We’re still shipping. Here’s what landed in the last few weeks. | SQLite allows one writer at a time. Turso 0.8 removes that limit with BEGIN CONCURRENT, and the numbers show it: • 500x lower tail latency. p99.9 of 2.4 ms at 32 connections, compared with 1.2 s for SQLite. • 7x higher write throughput. 9,500 transactions per second at 64 connections, compared with 1,370 for SQLite. The release also adds complete support for window functions, recursive queries with WITH RECURSIVE, full-text search that works with MVCC, and a smarter query optimizer. The benchmarks are open source, so you can run them on your own hardware. $ curl -sSL tur.so/install | sh |
Read the release notes → | There are four new AWS regions: Sydney, São Paulo, Central Canada, and Stockholm. That brings Turso Cloud to ten regions across four continents. Before, an app in Sydney reached its database in Tokyo at about 109 ms per round trip. In-region, it’s about 1 ms. Teams with data residency requirements can now keep their data in-country. Want Singapore, Paris, Seoul, or N. California next? Tell us on Discord. See all regions → | Agents delete things. Now you can undelete them. On any Turso Cloud paid plan, you can restore a deleted database for up to 5 days at no extra cost. Restore from the dashboard, or ask your AI coding tool to undo its own mistake through the Turso MCP server. Learn about Accident Protection → | // From the engineering team | | In v0.7, every transaction paid for its own fsync, so concurrent writers queued behind each other’s disk flushes. Pere Diaz Bou, Senior Software Engineer, explains how group commit lets a batch of transactions share a single fsync, cutting tail latency by three to five orders of magnitude. He also explains why Turso gets faster as you add connections. Want the rest of the story? Pere is speaking about breaking SQLite’s single-writer bottleneck at P99 CONF on October 21–22. Read the deep dive → | // Thank you | | To our customers, contributors, and community: none of this happens without you. Questions about today’s news? Reply to this email or find us on Discord. We’ll be around. The Turso team | | | Unsubscribe |
|
|
|
|
|