What types of AI proctoring does Tarkflo support?

Depending on the round and enabled capabilities, the product includes browser activity, camera-presence, gaze and head-pose signals, visible-object and display checks, supported audio signals, room scans, a paired side camera, cross-round identity context and answer-authenticity insights. The page separates these checks because they do not all run in every interview.

Are the same checks available in screening and live interviews?

No. Asynchronous video screening uses a narrower face, gaze and head-pose path. Live two-way interviews can support additional browser, object, display, audio and environment checks, depending on policy and workspace enablement.

Can we record flags without interrupting the candidate?

Yes. Watch-and-flag mode records available signals for the report without interrupting the candidate. Guide mode can pause for supported issues without automatically ending the interview. Only strict mode can end the round under its configured strike rule.

Do flags automatically reject a candidate?

No. A flag is a reason to inspect the evidence, not a finding of misconduct. An explicitly configured strict policy can end an interview after qualifying violations, but the hiring decision remains with your team.

Do room scans and dual-camera checks run in every interview?

No. These are configured two-way interview capabilities. The paired phone side camera additionally requires the relevant module. Its captured-frame trail and analysis should not be described as guaranteed continuous second-camera video recording.

Does an extra-display check mean the candidate’s screen is recorded?

No. Supported connected-display checks and camera-visible extra-screen signals are not screen sharing or screen-content recording. Browser support and permissions affect what can be observed.

Are identity and answer-authenticity insights definitive?

No. Photo comparisons depend on enabled settings, available photos and analysis providers. Answer-authenticity insights are probabilistic and depend on sufficient interview evidence. Missing or unavailable analysis is not a failed check, and neither output establishes misconduct by itself.

Can we control how long evidence is kept?

Raw integrity-event retention is managed separately from recordings, reports and hiring decisions. Confirm your workspace’s enabled settings and handling requirements rather than assuming every evidence type is deleted together.

AI Proctoring

AI proctoring. Context for human review.

Connect configurable interview checks to the moments your team needs to review. Choose how a round responds, inspect the evidence and keep the final hiring decision human.

Schedule a demo Explore the capabilities

For hiring teams that need interview-integrity context without treating every flag as a verdict.

A realistic remote-interview setup with a silver laptop, a mounted USB webcam, headphones and a smartphone on a desk stand.

Part of Tarkflo Hire

Detect a signal. Understand the context. Then decide.

A tab switch, another voice or a camera interruption can have more than one explanation. Tarkflo connects available integrity signals to the interview timeline, so reviewers can investigate what happened rather than rely on a single warning or an unexplained score.

See the connected hiring platform

Checks that fit the round

Configure monitoring and response behavior around your interview policy, enabled modules and the candidate experience.

A trail your team can review

Return to timestamped events, available recordings and supporting evidence instead of interpreting an isolated alert.

Human judgement stays central

Integrity flags and authenticity estimates support review. They do not establish misconduct or make the final hiring decision.

What is included

A layered view of interview integrity.

Coverage depends on the round, policy, browser, permissions and enabled workspace capabilities. Live two-way interviews support a broader set of checks than asynchronous video screening.

Browser activity

Context about how the interview tab is used, where supported and configured.

Tab focus and time away from the interview
Fullscreen exits and returns
Clipboard activity and developer-tools shortcuts

Camera presence

Camera-based geometry signals help reviewers understand visibility and interruptions.

Face leaving or returning to the frame
Persistent multiple-face presence
Camera, lighting and model-quality context

Gaze, head pose and gestures

Sustained or repeated movement patterns can be recorded as context, not proof of intent.

Sustained off-screen gaze and repeated glance patterns
Head turns and tilts
Sustained hand presence in the upper frame

Visible objects and extra displays

Supported live checks combine browser display information with camera-visible objects.

Visible phones, additional people and extra screens
Connected-display checks on supported browsers
No promise of screen sharing or screen-content recording

Audio and additional voices

Audio signals can highlight moments that need a closer listen in a two-way interview.

Possible additional-voice signals during the live round
Recording-based speaker segments where available
Review cues, not voice authentication or a lie detector

Room scan and paired side camera

Optional two-way interview checks add context about the candidate’s surroundings.

Guided room scan where configured
Review of visible people, screens, devices and notes
Paired phone side camera with a separate enabled module

Identity consistency across rounds

Where enabled and start photos are available, compare the person appearing in different rounds.

Cross-round start-photo comparison
Match, review, mismatch and unavailable states
Human comparison and resolution, not automatic rejection

Answer-authenticity context

Available analysis can surface response patterns for the team to investigate.

Per-answer signals and explanatory context
Overall insight when enough answers are available
Probabilistic review aid, not a definitive AI detector

Policy and evidence controls

Separate collecting a signal from interrupting or ending a round.

Off, watch-and-flag, guide and strict modes
Timestamped flags and available recording key moments
Raw-event retention settings separate from recordings and reports

01 / Set the interview policy

Choose the checks. Decide how the round responds.

Use workspace defaults and supported round settings to configure monitoring. Watch and flag without interruption, guide a candidate back when a supported issue occurs, or use an explicit strict policy with a configured ending rule.

Separate detection layers from response modes
Only strict mode can automatically end a round
Configured strike families and limits, not an automatic hiring rejection

02 / See the moments that matter

Review the event. Not just the alert.

Inspect timestamped browser, camera and available audio signals. Where the round was recorded, key moments can take a reviewer back to the relevant part of the recording. Missing or unavailable evidence stays visible as a limitation.

Key moments and individual event times
Available recording, room-scan and side-camera evidence
Camera quality and availability matter when interpreting signals

03 / Keep review grounded

Keep integrity context beside the candidate evidence.

Review policy interventions and available identity or answer-authenticity context alongside the interview. Look for an explanation, check what the evidence actually supports and record the human decision in the connected hiring workflow.

Pause and policy-ending history kept in context
Identity and authenticity insights where available
Final decisions remain with the hiring team

From setup to next step

How it fits into your workflow.

Configure the round
Choose the monitoring policy and supported checks, and confirm which optional modules and advanced features are enabled.
Prepare the candidate
Present the relevant monitoring information and required preparation, permissions and any configured room or side-camera steps.
Investigate and review
Open the timestamped evidence, consider technical or environmental explanations and decide the appropriate follow-up.

A little more clarity

Questions about AI Proctoring.

Better together

Career Page Builder

Build a branded careers experience with editable content, live job listings and applications connected to your Tarkflo hiring pipeline.

Explore the module FitScore CV Screening

Explore all products

Your workflow. Your next hire.

Make the next integrity review more informed.

Walk through the checks, candidate experience and evidence your team can actually use.

Schedule a demo

A practical walkthrough, built around your hiring process.

AI proctoring. Context for human review.