For embedding modelsCore Engine, <1 ms in-process

For serving at the edgeServe Engine, your machines & network

For building & operatingAgent Studio, bots & polyglot IDE

JavaKotlinScalaPythonONNXSelf-hosted
01

Core Engine

Sub-millisecond in-process ML, NLP, time series, anomaly detection, and deep learning for the JVM. Zero Python CVE surface, built-in TreeSHAP and KernelSHAP, and a perpetual OEM license for proprietary products.

See Core Engine →
02

Serve Engine

Put the model where the decision is. Score tabular models in <2 ms with <5 ms local TreeSHAP explanations in a 0.2 GB micro binary, or serve ONNX and local LLMs on your own servers without AGPL obligations.

See Serve Engine →
03

Agent Studio

The local-first AI agent workspace, 24/7 autonomous Bots, and polyglot IDE. Nine specialist agents, human-in-the-loop draft queues, and zero-code custom agent extensibility. Free to evaluate and bundled with commercial tiers.

See Agent Studio →

SMILE now seems to be the go-to general-purpose machine learning library for the Java and Scala worlds.

KDnuggets

A broad range of algorithms out of the box — and all you need is Java, or any JVM language.

InfoWorld

One of the hidden gems in the Java framework ecosystem today.

Patrick Martin, Citi

Latest

HALO.AI LLCNew York