Chrome extension

Is this page telling you the truth?

SignalCheck combines domain reputation with AI-content detection into a single 0–100 trust score. One click, two signals, about six seconds.

Install the extension See how the score works

Free. No account. Nothing about you is stored beyond 24 hours.

Two signals, not one

Most AI detectors read text in a vacuum. Whether a page is trustworthy also depends on where it came from — so SignalCheck checks both and tells you which one moved the score.

🌐

Where it came from

The URL is checked against Google Safe Browsing for malware, phishing, and unwanted software. A flagged domain drops the score hard and shows the threat type.

✍️

How it was written

The first 2,000 characters are analyzed by Gemini 3.6 Flash for signs of machine generation, returning a probability and the specific signals it noticed.

📊

One number, explained

The two signals combine into a single 0–100 score — and the popup always shows the parts, so you can disagree with the summary.

Reading the score

The score is a signal, not a verdict. It tells you how much scrutiny a page deserves.

70–100
High trust

Safe domain, content reads as human-written.

40–69
Medium trust

Mixed signals. Usually worth reading the breakdown.

0–39
Low trust

Flagged domain, strongly machine-generated content, or both.

score = (domain_reputation × 0.4) + ((100 − ai_probability) × 0.6)
The 40/60 weighting is a starting judgment, not a fitted parameter. Calibrating it against labeled data is what the current evaluation work is for.

Built in the open

Most products don't publish what they cost to run or how fast they actually are. These are measured from the live service, not estimated.

$0.0066
Inference cost per scan
~6s
Typical scan time
88%
Of cost is model output
24h
Before your data is deleted
Cost and latency are read from the model API per request and shown on a live admin dashboard. The full engineering write-up — including the code review that found three governance controls not enforcing what their documentation claimed — is in the case study.

Privacy by construction

The product needs to identify repeat users for rate limiting and nothing else. Everything follows from that.

✓
No accountsInstall and scan. There is no sign-up, because there is nothing to sign into.
✓
Your IP is never storedIt is hashed with a secret salt before it touches the database, and the hash cannot be reversed.
✓
Truncated in your browserOnly the first 2,000 characters ever leave your machine — the trimming happens before the request is sent.
✓
Deleted after 24 hoursRaw scan records are purged daily. Only anonymous totals survive.

What it can't do yet

A tool about trust should be honest about its own. These are open, not hidden.

!
Accuracy is being measured nowThe AI-detection layer is under evaluation against a labeled dataset. Until that lands, treat the score as a prompt to look closer rather than a judgment.
!
The weighting is a judgment call40% domain, 60% content was chosen by reasoning, not fitted to data. Calibration is planned.
!
Page text includes navigationExtraction takes visible page text, so menus and boilerplate can dilute the analysis on some sites.
!
One scan per dayA deliberate cost guardrail while the service runs on a small budget, not a paywall.

Install it

Chrome Web Store submission is planned once the V3 evaluation work is complete. Until then it installs from source in about a minute.

Clone the repository.
git clone https://github.com/meghaldonde/trustsignalcheck
Open chrome://extensions and turn on Developer mode (top right).
Click Load unpacked and select the extension folder.
Pin SignalCheck to your toolbar, open any article, and click Scan Page.
Coming to the Chrome Web Store after V3. The remaining work is accuracy evaluation — measuring how often the detector is right before asking people to rely on it.