start here

What do you want to inspect?
Choose the surface that matches the job.

SIGNALWATCH is easier to use when you enter with intent: monitor live signals, inspect evidence, test perception, or learn the ideas first. Nothing here asks you to trust a black box without a trace.

real source

Claims need source records or ingestion events.

model output only

Confidence and detections come from inference.

unavailable is valid

Missing data stays visible instead of guessed.

no prefilled claims

Protocols do not ship with conclusions.

open console

Monitor live AI signals

Open source-backed updates, collector health, and runtime state.

continue
open ledger

Inspect provenance

Review source claims, runtime frames, telemetry snapshots, and unavailable states.

continue
run lab

Test perception robustness

Upload an image, apply degradation, run real browser-side detection, and export a packet.

continue
review safety

Review safety context

Read risks, alignment concepts, and policy references with sources attached.

continue
learn terms

Understand AI basics

Start with definitions before entering the operational surfaces.

continue
open systems

Check system boundaries

See what the runtime observes, what it derives, and what stays unavailable.

continue
how to read the evidence labels
plain language / trust boundary
Real

Observed or sourced

A source link, timestamp, runtime event, or model output exists. The interface can point back to where it came from.

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.

guided route for first-time users
1. Learn the terms

Start with plain definitions for LLMs, confidence, telemetry, provenance, robustness, and evaluation.

2. Understand training

See how language models move from raw data to deployment monitoring.

3. Understand testing

Learn why AI systems are tested for failures, not only high scores.

4. Try perception

Upload an image or use a webcam to see real model outputs under degradation.

5. Run protocols

Use repeatable case-study protocols and export evidence packets.

what to trust
SIGNALWATCH does not invent detections, confidence, incidents, or source claims.
If a model emits no output, the interface shows that absence directly.
Source-backed AI updates keep links, timestamps, and provenance attached.
Educational pages explain concepts; they do not claim to reveal a private lab recipe.
inspect evidence ledger
start-page evidence boundary
real inputs / real outputs / explicit unavailable states
source data

papers, releases, policy pages, ingested records

traceable
model behavior

browser-side detections, confidence, timing, empty frames

observed
evidence packet

timestamps, frame history, continuity markers, exported JSON

recorded
claim boundary

missing data stays unavailable; outcomes are not prewritten

enforced
real source

Claims need source records or ingestion events.

model output only

Confidence and detections come from inference.

unavailable is valid

Missing data stays visible instead of guessed.

no prefilled claims

Protocols do not ship with conclusions.

if you are new to AI safety

What SIGNALWATCH is

Read the short trust boundary before entering technical surfaces.

Learn the words

Plain definitions for confidence, telemetry, provenance, and robustness.

See a real model fail

Use image degradation to inspect browser-side model outputs.

real-world visual context
illustrative photos / not model evidence

These images show the kinds of real-world conditions SIGNALWATCH is designed to explain: low light, camera deployment, motion blur, sensor boundaries, and human monitoring. They are source-attributed visual context, not detections or evaluation results.

visual role
operational context
source status
externally attributed
evidence boundary
not inference output