The platform has a variety of features which can be applied while conducting an online exam, to prevent cheating by candidates during the exam. 

1. Access Time: Allowing the candidate to take the assessment(s) only during a designated/pre-decided time slot. The candidate will not be able to start the test outside of the access time set for the test.

2. Mettl Secure Browser (MSB): This browser makes the candidate’s system into a safe lockdown mode which prohibits any third-party application in the background that can be used for cheating. For this the candidate is prompted to install the browser on their system and then proceed with examination.

3. Browsing tolerance: This feature counts every time a test-taker navigates away from the test window to give warnings (optional) if the set limit is exceeded the test will be terminated. The platform also provide a feature wherein there is an option to show the remaining count of the number of times the candidate navigates away from the test window.

Following are the features available under Browsing Tolerance:

• Not Applicable – After enabling this feature, Browsing Tolerance will not be applicable for that assessment. 

• No Warning – After enabling this feature, the test will be terminated without any warning if the candidate navigates outside the test window.

Note: We can select the warning counts from the options, or we can also select the Custom-warning limits as per the requirement of the test.

4. Advanced Visual Proctoring: After enabling this feature, the candidate has to submit the snapshot of their ID card and their face, it also has a feature to “Capture Candidates’ Screen”.

Following are the flags generated by the AI:

• Capture Candidate Screen – This feature captures the candidate’s screen & Camera feed snapshots at every 60 seconds. Also, along with the reports, we can view the candidate screen snapshots while viewing the live feed of a candidate too which captures the face of the candidate at every 60 seconds.

• Candidate Authorization - Candidate authorization is a manual intervention, and the authorizer can allow/block the candidate which will confirm candidate’s authenticity. 

• Store Recorded Video – This feature allows us to capture candidate's entire webcam input for the entire test which can be previewed in the Test-taker’s report.

• Audio Proctoring - While scheduling a proctored test, we can also listen to the audio feed of the candidate while they attempt the test. This feature will detect any human speech and the number of speakers during the test. By default, this option is disabled for all the tests. 

• Mandatory 360 Check – This prompts the candidate to show the entire room before taking the test through their webcam.

5. OTP (One Time Password): Protected Test-takers can access the test link only after entering the One-Time Password, sent to their registered email address. However, if Mettl Secure Browser is enabled, OTP will not function.

6. IP Address Restrictions: This feature allows test-takers to attempt the test from a defined IP range. No other range of IP addresses would be allowed to access the test link. Only the systems on the range specified will be able to successfully take the test. The IP address must be a Public IP address (static) and not Private IP address.

7. Private Access: Allowing only registered candidates to take the test, through verified test-registration fields. This makes sure that only desired candidates are able to access and take the test. There is also an option to send customized invitation email notifications to the candidates with the protected link to the test. Click here to know more.

8. Supervised resume: Allow only genuine candidates to resume a disconnected test in case of network failure or other technical issues. A proctor is required to do the same.

9. No Plagiarism: Blocking "copy-paste", "print-screen" controls.

10. Watermark: This feature displays a watermark in the background of the test screen to discourage candidates from cheating via taking screenshots or pictures of the test window. It recognizes the point of leak to control exposure of content.

11. Question and Option Randomization: This allows randomizing all the questions and options shown to the test taker to discourage them from cheating. One option is to randomize the options and questions with the same skill set and the other is to randomize amongst multiple skill sets. Skill sets are nothing but topics containing multiple questions.

12. Detecting code generated via ChatGPT (Generative AI Plagiarism): To combat the heighten concerns around how one can prevent selecting candidates who could have possibly taken help of Generative AI tools while attempting tests, we have strengthened our reporting with an additional code quality indicator, namely AI Sense score. AI Sense is an in-house algorithm that harnesses the power of generative artificial intelligence (AI) to sense and identify instances of AI-generated code.