Skip to content

// use case · Founders, growth engineers and billing owners at SaaS, AI and API products with a time-limited or credit-based trial.

Stop people farming your free trial with throwaway inboxes

Most trial farming relies on fresh inboxes, so the email is the cheapest place to stop it. Look up each trial signup, refuse the trial when verdict.is_spam is true, and keep webmail users in the normal flow. Combine the email result with your own signals (payment fingerprint, IP, usage) for repeat offenders.

The problem

  • The same person has had our 14-day trial a dozen times with different addresses.
  • Our AI credits are getting burned by accounts that never convert.
  • I do not want to force a credit card on everyone just to stop a few abusers.
  • Blocking all Gmail signups would kill real conversions.

How pingutil solves it

  1. 1

    Check the email before granting the trial

    Create the account if you like, but call /v1/lookup with skip_enrichment: true before you attach trial credits or start the clock.

  2. 2

    Withhold the trial on is_spam

    Disposable domains and randomized local parts set verdict.is_spam. Ask for a permanent address before the trial starts instead of silently failing.

  3. 3

    Treat webmail as normal

    category free means Gmail, Outlook and similar. That is a real person most of the time, so keep them in the standard trial and lean on your other signals.

  4. 4

    Key repeat checks on the registrable domain

    The domain field is the eTLD+1, so subdomain tricks on custom domains collapse to one value you can count against.

Fields you will use

Response fieldWhat to do with it
verdict.is_spamWithhold trial credits
verdict.reasonSeparate disposable_domain from randomized_local_part in your logs
email_signals.randomness_scoreFeed a 0 to 100 score into your own risk model
categoryKeep free webmail users on the normal path
domainCount trials per registrable domain
import os, requests

res = requests.post(
    "https://pingutil.com/v1/lookup",
    headers={"Authorization": f"Bearer {os.environ['PINGUTIL_KEY']}"},
    json={"input": email, "skip_enrichment": True},
    timeout=5,
).json()

if res["verdict"]["is_spam"]:
    return deny_trial(reason=res["verdict"]["reason"])

risk = res["email_signals"]["randomness_score"]
start_trial(user, risk_hint=risk)

Compared with the alternatives

Common options are a credit card wall, phone verification or a full fraud platform. A card wall stops most farming but also stops a lot of honest evaluators; phone checks add cost per user and friction. Email signals are cheaper and invisible to real users, but a determined abuser with real Gmail accounts gets past them, so they work best as the first layer.

See provider comparisons →

Limits to plan for

  • Real Gmail or Outlook accounts are not spam by any email signal, so a patient abuser with many webmail accounts needs your own device or payment checks.
  • Plus-addressing and Gmail dots are not merged for you; normalize those yourself before counting repeats.
  • Some real non-English names can score as random, so use randomness_score as a hint, not a ban.

Common questions

Is requiring a credit card the better fix?
It stops more abuse but costs you honest evaluators. Many teams check the email first and only ask for a card when signals look risky.
How do I catch the same person using plus addresses?
Normalize the local part yourself: strip the plus tag and, for Gmail, the dots. Then count trials per normalized address and per domain.
Will this block people using Apple Hide My Email?
Relay addresses can be flagged as randomized. If that matters for your audience, allow privaterelay.appleid.com and similar relay domains before acting on randomized_local_part.

Read next