Technical deep dives on Tokio backends, copy-on-write storage, and the future of PostgreSQL.
Every app, every agent, every workflow needs a backend. We're making them instant, disposable, and free when idle.
Containers made compute disposable. It's time to do the same for backends. Here's why — and how Tokio makes it possible.
Legacy PostgreSQL spawns one OS process per connection. Trubase DB replaces that with Tokio async tasks weighing kilobytes.
How data lives as immutable layers on object storage, and how branching a backend is a single metadata write.
A backend that weighs kilobytes changes what one machine means. The task math, the overbooking economics.
Not 'minimum instance.' The tasks drop, the memory is reclaimed, the state persists as cold bytes. $0.
StartupMessage, SimpleQuery, ExtendedQuery pipeline — what implementing PostgreSQL v3 wire protocol from scratch involves.
Every AI agent invocation needs isolated state. Branch from a template in milliseconds. Mutate freely. Dispose when done.
Every hosted Postgres platform wraps the same legacy C engine. Here's why that makes backend-as-a-value impossible.
Rebuilding Postgres sounds unverifiable — until you realize Postgres ships its own proof. The scoreboard, not the resume.