AI shouldn't cost the earth.
We build tools that make artificial intelligence safer to use and cheaper to run — protecting people from harm and the planet from waste.
Our approach
Deterministic over probabilistic. Every check produces the same output every time. No model to prompt-inject. No hallucinations. No distribution shift. A Rete forward-chaining network replaces the LLM judge — same security boundary, zero probability.
Efficiency over scale. Our inference engine runs at 124µs per check on a single core using ~50 MB of memory. That’s roughly ten thousand times more energy-efficient than calling an LLM for the same decision.
Ungameable by design. Actions are parsed into ASTs and content-addressed before classification. Obfuscation, encoding tricks, and prompt injection don’t work. Ambiguity defaults to deny.
Products
ellm-guardrail
Deterministic policy enforcement for programmatic tool execution. Define dangerous actions in TOML. Block them in 124µs.
Supply-chain safety
Deterministic policy checks for CI/CD pipelines. Prevent poisoned dependencies from reaching production.
From the journal
- July 2026
Why deterministic beats probabilistic for security boundaries
When the cost of being wrong is catastrophic, 99.9% accuracy isn't good enough. Here's why we bet on forward-chaining inference instead of LLM judges.
- July 2026
The carbon cost of "just ask the LLM"
Every unnecessary model invocation burns electricity. We measured how much — and built an alternative that runs on a Raspberry Pi.