Real
Observed or sourced
A source link, timestamp, runtime event, or model output exists. The interface can point back to where it came from.
about / trust boundary
SIGNALWATCH is an evidence-aware observability surface.
It helps readers inspect AI safety signals, source movement, perception robustness, and runtime state without filling gaps with invented telemetry.
public route
Live console
A realtime operational surface for source-backed AI updates and runtime state.
open surface
Evidence ledger
A ledger of source claims, telemetry frames, collector state, and unavailable data.
open surface
Perception lab
Browser-side COCO-SSD inference under real input degradation.
open surface
Safety context
Source-backed safety concepts, frameworks, and public references.
open surface
the four labels that make the site readable
plain language / trust boundary
Derived
Calculated from real inputs
A summary, grouping, or trace built from source activity, telemetry, or model-output history. It should still show its inputs.
Conceptual
Explanation, not measurement
A teaching example that explains a risk or system behavior without claiming it happened in a deployed system.
Simulated
Controlled demonstration
A parameter-driven demo. Useful for learning, but not evidence about the outside world unless real inputs are attached.
what this is
SIGNALWATCH is a public, operational interface for inspecting evidence around AI systems. It combines source monitoring, runtime telemetry, safety references, and browser-side robustness tests while keeping each data type labeled.
The goal is not to dramatize AI. The goal is to make evidence easier to read: what was observed, where it came from, when it appeared, and what remains unknown.
what this is not
A synthetic intelligence simulator.
A startup-style analytics dashboard with decorative metrics.
A claim that a model, company, or deployment is safe.
A source of fabricated incidents, confidence values, or telemetry.
trust boundary
source-backed / model-reported / unavailable stays visible
Source claims
Must include real source data, links, timestamps, or provenance.
Runtime telemetry
Describes SIGNALWATCH infrastructure state, not facts about the AI ecosystem.
Model outputs
Come from real browser-side inference. Missing outputs remain missing.
Conceptual demos
Are labeled education surfaces, not operational measurements.