Skip to content

Embedded & Serverless

Uni runs as a library inside your application process. There is no separate server to deploy, monitor, or scale. You open a database path and start querying immediately.

What It Provides

  • In-process execution with no network hops.
  • Simple local deployment for desktop, backend services, or edge devices.
  • Optional durability to local disk or object storage.

Example

use uni_db::Uni;

#[tokio::main]
async fn main() -> Result<(), uni_db::UniError> {
    let db = Uni::open("./my_db").build().await?;
    let rows = db.query("MATCH (n) RETURN count(n) as c").await?;
    println!("{:?}", rows);
    Ok(())
}

import uni_db

db = uni_db.Database("./my_db")
rows = db.query("MATCH (n) RETURN count(n) AS c")
print(rows)

Use Cases

  • Local search and recommendation inside a service.
  • Edge or on-device knowledge graphs.
  • Testing and CI environments without external dependencies.

When To Use

Choose embedded mode when you want predictable latency and minimal ops. If you need multi-tenant hosted access or cross-region write scaling, use a separate managed system.