Tako is a browser-local atomistic simulation workspace for quantum chemistry. Tako included a native jupyter notebook like interface and rich agentic support, making quantum chemistry more approachable. The underlaying chemistry Atomli reduced DFT compute by more then 10x without any compromise on accuracy, reduce compute cost.

What became possible in your product with Astra that was not practical before?

Astra has deep understanding in linear algebra and way much better Rust coding ability, allowed me to implement many optimizations fast.

Quantum chemistry software has remained heavy, fragmented, and difficult to deploy. Most packages depend on complex toolchains, large runtime environments, and loosely integrated components. A truly lightweight, vertically integrated stack has been missing. We are rebuilding that stack from the foundations. By reimplementing the core machinery of Density Functional Theory in Rust, we achieve an exceptionally small footprint and extremely high evaluation speed, without sacrificing numerical accuracy. The architecture is designed around modern hardware from the start: through WGPU, the same portable runtime extends across CPUs and GPUs, with much of the computational pipeline available directly on the GPU. And we are building for where computational chemistry is going next. We believe machine-learned functionals and interatomic potentials will become a core part of the computational chemistry stack. That is why Skala XC and multiple MLIPs are supported from day one—not as external add-ons, but as first-class components of the platform. Lightweight. GPU-native. ML-ready. A modern computational chemistry stack, rebuilt from first principles.

What became possible in your product with Astra that was not practical before?

Astra has deep understanding in linear algebra and way much better Rust coding ability, allowed me to implement many optimizations fast.

Tako by Atomli

Fastest, lightest quantum chemistry simulation.

Fastest, lightest quantum chemistry simulation.