Spam protection

How FormDB protects forms from spam: the honeypot, the eight content classifiers, strict and relaxed modes, and how the Spam folder works.

This page gives an overview of FormDB's spam protection: the layers available, how the honeypot and the spam filter work, and what happens to submissions that are flagged.

The layers

FormDB has five layers. They run in this order, and you can combine them.

Layer

What it does

Result of a match

Plans

Allowed domains

Accepts browser submissions only from websites you list

Rejected

All plans

Blocked IPs

Rejects specific IP addresses or ranges

Rejected

All plans

Honeypot

Catches bots that fill in a hidden field

Silently dropped

All plans, always on

Captcha

Requires a solved built-in captcha, or a token from reCAPTCHA, hCaptcha or Turnstile

Rejected

All plans

Spam filter

Scans the content for spam patterns and checks your own spam rules

Stored in the Spam folder

All plans

For most small sites, the honeypot plus the spam filter is a good start. Add a captcha if bots still get through.

Honeypot

A honeypot is a field that is hidden from people but visible to bots reading your HTML. Many bots fill in every field they find. A filled-in honeypot gives them away.

FormDB's honeypot field is named _gotcha. Add it to your form:

<input type="text" name="_gotcha" tabindex="-1" autocomplete="off" style="display:none" />

When _gotcha arrives with any value:

  • The bot receives the normal success response, so it does not learn that it was caught.
  • Nothing is stored, not even in the Spam folder.
  • The submission does not count toward your quota.

The honeypot is always on. There is nothing to enable. The HTML, React and Vue code generated by FormDB already contains it. If you write your own form, or use the Fetch or jQuery snippets, add the field yourself.

Spam filter

The spam filter scans the values of each submission for words, phrases and links that are common in spam. It is off by default.

Turn it on

  1. Open the form and go to the Settings tab.
  2. In the Spam filter card, turn on Enable spam filter. All eight classifiers are selected for you.
  3. Choose a Sensitivity and untick any classifier you do not want under Detect.
  4. Click Save.

The eight classifiers

A classifier is one category of spam the filter looks for.

Classifier

What it looks for

Profanity

Offensive language

Crypto

Cryptocurrency and wallet pitches, for example "bitcoin", "airdrop", "seed phrase"

Drugs

Drug sales and references

Fraud

Scams, prizes and wire transfers, for example "lottery winner", "verify your account"

Adult content

Explicit or adult services

Spammy phrases

Sales phrases such as "act now", "free money", "click here", "seo services"

Suspicious links

Link shorteners such as bit.ly and tinyurl.com, and links on frequently abused domain endings

Disposable emails

Addresses from throwaway inbox providers such as mailinator and yopmail

How the filter works:

  • It matches fixed lists of keywords and patterns, ignoring upper and lower case. It is not based on machine learning and does not use IP reputation.
  • It scans the submitted values, not the field names.
  • If the filter is on but no classifier is ticked, nothing is flagged.

Your own rules

The Spam filter card also has rules you set yourself. A submission that breaks one of them is flagged as spam:

  • Blocked words or phrases, matched without regard to upper and lower case.
  • A maximum number of links in the submission.
  • Blocked countries, or allowed countries so that every other country is flagged.
  • Allowed languages, so that messages detected in another language are flagged.
  • Duplicates: the same content sent to the same form within 24 hours.

Because it matches keywords, the filter can flag a genuine message. A shop that sells diet products, for example, should untick Spammy phrases or use Relaxed. Check the Spam folder now and then, especially after turning the filter on.

Strict and relaxed

Sensitivity

A submission is spam when

Strict (default)

Any single classifier matches.

Relaxed

Two or more classifiers match, or the matching classifiers make up at least half of the ones you enabled.

With only one or two classifiers enabled, Relaxed behaves the same as Strict.

What happens to flagged submissions

A submission flagged by the spam filter:

  • Is stored in the form's Spam folder, not the Inbox.
  • Does not count toward your monthly quota. It is stored even when your quota is used up.
  • Does not send a notification email.
  • Is not sent to webhooks or other integrations.
  • Loses its uploaded files. Files of spam are not stored.
  • Looks successful to the sender, who gets the normal success response.

The Spam folder

Open the form's Submissions tab and click Spam.

  • Click a submission to open it. The Details section shows Spam signals: which classifier matched and the text that triggered it.
  • To rescue a genuine message, select it and click Not spam. It moves to the Inbox.
  • To flag something the filter missed, select it in the Inbox and click Spam.

Moving a submission out of Spam does not send its notification email or run integrations after the fact, and it does not use quota. Marking a submission as spam by hand does not give quota back.

Spam is kept for as long as other submissions on your plan. See Data retention.

The dashboard and each form's Analytics tab show how many submissions were caught as spam.

What FormDB does not do

  • There is no rate limiting per visitor or per form.
  • FormDB does not use Akismet or other outside spam services.

Next steps

Ready to try it? Create a free account and send your first submission.