Skip to main content

Our Mission

Our mission is to build resilient, secure, safe, deterministic, and scalable AI-native solutions that unlock developers' full AI momentum with a design target under 2 ms — or die trying.

We are aggressive about making SigmaShake fast, trusted, and reliable with no compromises. We do not watch the competition; we focus on building great products.

What this means in practice​

  • Resilient — the engine fails closed, survives daemon crashes, and recovers without operator intervention.
  • Secure — every tool call is evaluated at the boundary; no LLM sits in the enforcement loop by default. (If you turn on the optional Jev assist with your own TypeSafe key, ssg eval sends the tool name and input to api.typesafe.ai, and an Ask result can override a local Allow.)
  • Safe — rules are declarative and auditable; a DENY or ASK rule stops or pauses a call; calls that match no rule are allowed; Hub rulesets are Ed25519-signed.
  • Deterministic — the same input always produces the same verdict. No drift, no hallucination.
  • Scalable — pure-Go evaluation, fleet-wide sync, and SSO for teams of any size.
  • Fast — the enforcement engine is pure Go, in-process, and its hot-path design target is under 2 ms (measured p50 0.39 ms, p99 1.91 ms in the 2026-06-10 sigmashake-bench daemon-socket run, ssg 0.29.156, n=500).