SYNTONY / SYSTEMS & SOFTWARERESEARCH → ENGINEERING

From evaluation
to control.

Tools for testing AI systems, governing their use, and examining the evidence. Built around the decisions that follow an evaluation.

01 / CONTAINMENT SYSTEMSIN DEVELOPMENT

CUTPLANE

Authority outside
the workload.

Independent authority for AI evaluation. CP-01 brings exact-request authorization, latched stop and signed evidence into a dedicated control-node design.

Explore CUTPLANE
CP-01 development rendering with schematic control interfaces
CP-01 / DEVELOPMENT RENDERING
Dedicated control-node architecture.
SCOPED AUTHORITYEXPLICIT STOP / REARMINSPECTABLE EVIDENCE
02 / SOFTWARE

The decision layer.

UNDERSTAND

Resonance

Inspect how risks interact. Compare a baseline with an illustrative intervention and examine the model behind the result.

INTERACTIVE RESEARCH CONCEPTOpen Resonance ↗
DECIDE

AgentProof

An AI investigation agent for acceptance evidence. Inspect supplied results, challenge measurement gaps, and preserve a review against a versioned record.

LOCAL DEMO AVAILABLE / HOSTED ACCESS OPENING SOONExplore AgentProof ↗
MONITOR

LiveMonitor

A direction for post-deployment monitoring: connect observed system behavior to the decisions that need attention.

ROADMAP / SCOPE UNANNOUNCEDDiscuss a monitoring need ↗
03 / AVAILABLE BY ENGAGEMENT

Start with
a real system.

Scope an evaluation, test the boundaries, and decide what the findings require. Syntony's evaluation, red-teaming and governance engagements are available while the product programs develop.

Explore our capabilities
04 / NEXT

2027

Enter the preview.

Development preview. Release scope and timing have not been announced.