SilentShield

Bot protection

Invisible bot detection for your forms

SilentShield watches how a form is filled in, not who fills it in. Real visitors pass without noticing; bots are stopped or asked to prove themselves — and you can see why.

The SilentShield dashboard with recorded form interactions, each with its verdict human, suspicious or bot
Behaviour logs: every interaction with its verdict and score. Screenshot of the English dashboard.

SilentShield's invisible bot detection scores every form interaction from 0 (bot) to 1 (human) based on mouse, keyboard, scroll and touch behaviour and the time the form takes. The result is a verdict — human, suspicious or bot — and you decide per verdict and per form whether to let it through, ask the visitor or reject it.

From the first click to the verdict

  1. A visitor fills in the form

    The widget measures how: mouse movement, typing rhythm, scrolling and touch, and how long the form takes.

  2. Your server asks SilentShield

    When the form is submitted, your server or the plugin checks the result through the API — the visitor notices nothing.

  3. SilentShield returns a verdict

    Human, suspicious or bot, based on a score from 0 to 1.

  4. You decide what happens

    Let it through, ask the visitor or reject it — set per verdict and, if you like, per form.

Submissions without any measurement, for example with JavaScript switched off, cannot be asked anything: for these you choose between rejecting and letting through.

What it gets you

No puzzles for real visitors

In the default mode a visitor is only asked when the behaviour looks off. Everyone else passes without any friction.

A reason for every block

Blocked and suspicious requests carry reason codes — typing too fast for a human, a perfectly straight mouse path, no pauses between words — so you can see why instead of guessing.

You set how strict it is

Ask only when something looks off, always (for sign-in or payment), or never. Individual forms can differ from the rest of your key.

Try it without risk

With the mode set to never and the actions set to let through, SilentShield scores and reports every submission but enforces nothing — the honest way to start in monitoring mode.

Checks that match the doubt

Proof of work

A small computing task that runs in the background. A person submitting one form barely notices it; a bot submitting thousands of forms is slowed down by every one of them.

Maths question as fallback

Only if the background task cannot be completed, a short, visible question appears — in 25 languages and with text-to-speech.

Honeypot fields

Hidden form fields that people never see and bots fill in. Switched on by default, no configuration needed.

Data protection

SilentShield sets no cookies and tracks nobody across sites. Browser and device characteristics contribute 10 % of the score and feed the fingerprint rules; if you would rather they did not, switch on Zero-PII mode — that removes this component entirely.

Questions and answers

Does a real visitor ever see a puzzle?

Only if the behaviour looks off. Then a computing task runs invisibly in the background first, and a short maths question appears only as a fallback. In the default mode, visitors with an unremarkable score are never asked.

Can I try it without blocking anyone?

Yes. Set the challenge mode to never and the actions to let through. SilentShield still scores and reports every submission, but enforces nothing.

How do I find out why a submission was blocked?

Every blocked or suspicious request is tagged with one or more reason codes, for example typing that was too fast for a human or a mouse path that is perfectly straight.

What happens if JavaScript is switched off?

Then there is no measurement for that submission, and nobody can be asked. You decide whether such submissions are rejected or let through.

Try invisible bot detection on your forms

Start free and see the verdicts for your own traffic.