What we believe

Three things we're betting on.

Compiled, not interpreted

The next generation of data and AI systems needs runtimes that compile to native code on the hot path. Bytecode interpreters and remote sandboxes lose. We design the opposite.

Open, not walled

Source-available, with no managed-service-only walls. You can self-host, fork, audit, embed. We don't ship products we wouldn't run ourselves.

Engineered to ship

A small Rust shop with tight scope, real benchmarks, and code that runs in production. No demoware, no buzzword stack, no half-finished pivots.

What we ship

Two products today. More coming.

ScramDB

Our database product. Adaptive HTAP, JIT-compiled SQL, PostgreSQL-compatible. Open source.

Visit scramdb.com for the architecture, benchmarks, and install guide.

Afterburner

Our JavaScript runtime for Rust. Capability-gated, embeddable, with the burn CLI. Open source.

Visit afterburner.sh for docs, examples, and the install one-liner.

Built on the right primitives

We don't reinvent fundamentals. We wire them together correctly. Compiled execution and capability-gated isolation, built on infrastructure that has been proven for decades.

Rust
LLVM
WebAssembly
Wasmtime
PostgreSQL
Node.js
WASI
JIT
Open Source
Berlin
Self-host
BYOC

Berlin. Rust. Since 2023.

We're a small team. Founded in Berlin in 2023. We write Rust because we care about what Rust gets right: zero-cost abstractions, no GC pauses, memory you can reason about. We ship the kind of systems we'd want to depend on.

Run it your way

Self-host, BYOC, or use our managed service when it ships. Same binaries, same behavior, your choice. Data stays where you put it.

What's next

Managed surfaces for both products. More runtimes for more host languages. Compiled execution wherever computation lives. That's the long bet.

Integration

Integrate seamlessly

icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
icon
Amazon S3
Apache Cassandra
Apache Kafka
Azure Blob Storage
Elastic
Google Storage
MongoDB
MySQL
PostgreSQL
Redis
SAP
Snowflake
Amazon S3
Apache Cassandra
Apache Kafka
Azure Blob Storage
Elastic
Google Storage
MongoDB
MySQL
PostgreSQL
Redis
SAP
Snowflake
FAQ

Frequently Asked Questions

Both compile to native code, both are built by Psila.AI, and Afterburner powers user-defined functions inside ScramDB. They ship as independent products because they serve different needs: ScramDB replaces your database, Afterburner embeds into any host that needs safe, fast scripting.

We build the runtimes.
You ship the product.

Open source
Berlin-based
Built in Rust