Does Rupt score answer quality?

No. Your checks score the answers after a respondent is in. Rupt scores the respondent before they are in: a real device, unique across the panel, human, and where it says it is. A duplicate with perfect answers passes an attention check and fails a device check. Speeders, straightliners and open-end scoring stay with your data quality tools.

Will real respondents notice it?

No. There are no extra questions and no captcha, so interview length does not change. Only a borderline panel signup is asked for a code by email or SMS.

Our panelists share IPs. Universities, hospitals, offices.

An IP alone is one signal, not a verdict. Rupt reads the device, the email, the phone and the history together, and your policy sets the response per country and per supplier. Twenty students on one campus network look like twenty devices, and one person on twenty accounts looks like one device.

Can it catch AI agents taking surveys?

Yes. Rupt classifies each session as human, declared agent or undeclared automation and names the framework where it can. Agent traffic is stopped before the first question, and you decide what to allow.

How does it fit into our survey flow?

Two places. At panel signup and login through the SDK or the API, and at survey entry with the JavaScript snippet on the page respondents land on. It runs independently of survey logic, so nothing in the questionnaire changes.

What data does it collect?

Device and network signals, plus the email address, phone number and IP you pass at signup so Rupt can check their reputation. Rupt never receives survey answers or panel profile fields.

Plans start at $99 a month with 20,000 evaluations included, then $0.005 per evaluation. An evaluation is one check at signup, login or survey entry, so cost follows the entries you check rather than your panel size.

Survey fraud detection for panels and market research

Duplicate respondents

One person, one panel account.

Link panel signups and survey entries by device, not by email, cookie or IP.
Survives cleared cookies, incognito, new emails and burner phones.
Catch professional respondents rotating identities across waves and panels.

Everything on the surface changes with every attempt. The device does not, and that is what Rupt links on.

Read the guide

Keep bots and AI agents out of your completes.

Classify every session as human, declared agent or undeclared automation.
Catch headless browsers, automation frameworks and agent traffic before the first question.
No captcha and no extra questions for real respondents.

Rupt takes a session apart down to the browser engine, the automation driver, the network and the timing between inputs. Automation has to get every one right to pass as human.

Read the guide

Know where a respondent really is.

Flag VPN, proxy, Tor and hosting IPs on every entry.
Corroborate location across a device's history, not a single IP.
Tune the response per country and per supplier, so a VPN alone is never a hard block.

City-level geolocation, impossible travel and four network flags on every session, returned in under a second.

Explore the signals

Stop farmed incentives at panel signup.

Score every panel signup on email, phone, network and device.
Catch disposable emails, numbers already verified on other accounts, and devices already on the panel.
Challenge borderline signups with a hosted email or SMS check.

A burner line passes every phone lookup as valid, active and mobile. Rupt links the accounts a number is verified on, and the device behind them.

Read the guide