THE PUBLIC HOUSE / 2026

AI makes claims.
Reality answers.

SYSTASYS makes tools that question what intelligent systems say they have done. What can be observed is shown. What cannot is a limit, not a promise.

Build.Check.Show.

S / 001 OBSERVABLE

OUR APPROACH

  1. 01
    The claimWhat was said
  2. 02
    The traceWhat can be accessed
  3. 03
    The checkWhat can be established
  4. 04
    The receiptWhat remains inspectable

A conceptual illustration, not an actual run.

METHOD / 04
INSPECTDISTINGUISHATTEST

01 / INTENT

Better decisions
start with evidence.

We don't ask anyone to take a polished demo on faith. We favor usable tools, inspectable sources, and a clear distinction between a fact, a simulation and an unknown.

02 / PRODUCTS

Three projects.
Three starting points.

Each tool retains its own identity, scope and limitations. Links lead to existing sources or websites, not a promise of outcomes.

01 / 03PUBLIC CODE · MIT

CODING AGENT VERIFICATION

SkeptOpus

Checks specific coding-agent claims against observable traces. Distinguishes evidence found, missing evidence, contradiction and inability to verify.

Inspect the source
02 / 03PUBLIC CODE · APACHE-2.0

EVIDENCE BEFORE ACCEPTANCE

ACE Receipts

A CLI and GitHub Action to inspect evidence, risk and permissions in AI-agent work before deciding whether to accept it.

Explore the repository
03 / 03PRODUCT WEBSITE

REVIEW BEFORE SENDING

Sendable?

A pre-send review experience that helps distinguish send, revise and don't send. Its own product website is available separately.

Discover Sendable?

03 / EVIDENCE & LIMITS

Evidence cannot fit
inside a tagline.

A public project is not a universal guarantee. To understand a result, inspect its source, context and limitations. When those aren't available, we don't make them up.

SOURCE
An inspectable original
SCOPE
What the tool actually checks
LIMIT
What remains unproven

04 / STORIES

See ideas
take shape.

Asso covers technology and research. Its voice and character are synthetic and disclosed as such. Those stories are separate from evidence about how our tools performed.

Watch Asso's stories