You write the code.
Shuttle writes the cloud.

This is the exact moment Shuttle was built for. Type one command, and the database appears. Annotate a function, and the secret store is provisioned. There is no Dockerfile to debug at midnight, no Terraform to memorize, no cloud console with four hundred buttons. The infrastructure is derived from the code itself - a philosophy the company calls, plainly, "infrastructure from code."

More than 20,000 developers have decided that is a better way to live. In October 2025, Y Combinator and a roster of unusually well-known angels agreed, writing a $6 million seed check.

The config file, quietly deleted

Most cloud platforms hand you knobs. Shuttle hands you back your afternoon. Here is what that looks like in practice.

// Ask for a Postgres database.
// Shuttle provisions it for you.
#[shuttle_runtime::main]
async fn main(
    #[shuttle_shared_db::Postgres]
    pool: PgPool,
) -> ShuttleAxum {
    let router = Router::new()
        .route("/", get(hello));
    Ok(router.into())
}
// $ shuttle deploy  ->  it's live.

From Rust tool to AI platform engineer

Shuttle spent years as the friendliest way to ship a Rust backend. Then the ground moved.

Founded in 2019 and forged in Y Combinator's Summer 2020 batch, Shuttle started narrow and deep: make deployment disappear for Rust developers. It worked. The open-source project on GitHub drew a devoted following, and the platform became one of the most popular ways to put a Rust backend online.

Then tools like Cursor, GitHub Copilot and Lovable arrived and made writing code almost trivial - and made the deployment gap glaringly obvious. AI could generate an app in seconds; it still could not make it run. Shuttle's bet is that the same engine which turns Rust annotations into infrastructure can become a language-agnostic "AI platform engineer" - one that assesses your app, presents infrastructure options with pricing, and provisions a database or buys hosting from a plain-English prompt.

Built for the person who just wants it live

Individual developers, small teams, and increasingly anyone shipping AI-generated apps who hit the same wall between "it works on my machine" and "it works."

Things worth knowing

• Shuttle is written 100% in Rust - the tool and the platform run on the same language they deploy.
• The whole pitch is "infrastructure from code": you never write a Dockerfile or a Terraform file.
• The company predates the AI coding wave by years - founded in 2019 - which made its "vibe-coding deployment" story newly, almost accidentally, timely.
• Its GitHub org, shuttle-hq, keeps an "awesome-shuttle" list of community projects built on the platform.
• The CEO traded his way through a quant hedge fund before concluding that the deploy button was the bigger fortune to build.

The wall, gone

Shuttle did not make the code better. It made the distance between "written" and "running" almost disappear - and it is betting the next version of that distance gets crossed by a prompt instead of a command. For 20,000 developers, that trade has already been made.