Why we built Clevid for the desktop

Over the past two years, engineering and university hiring teams noticed the same pattern: browser-based assessments stopped predicting live interview performance. Candidates were clearing online screening rounds with near-perfect scores, then struggling to reason through basic variations of their own submissions forty-eight hours later.

The root cause is simple. A web browser tab is isolated from the rest of the computer by design. When you run a high-stakes assessment inside a standard browser tab, you are asking an environment built for web browsing to verify what is happening across the rest of the workspace.

What a browser tab cannot see

Because a browser tab can only see inside its own webpage, it is blind to:

• anything displayed on a second monitor,
• any application window sitting outside the browser tab,
• transparent AI overlays that hide themselves from standard screen sharing,
• background screen-sharing streams to another person,
• whether the assessment is running inside a virtual machine while help runs on the main computer.

At the same time, there are things no software on the laptop can see: a phone propped up off-camera, a second person in the room, or notes beside the monitor. Designing a trustworthy assessment platform requires being honest about both lists.

The shape of the platform

Clevid is a proctored technical assessment platform built in two halves.

The HR console is a fast web workspace. Author coding and multiple-choice assessments, sample per-candidate papers from structured question banks, invite cohorts via CSV, and review ranked results. Recruiters and hiring managers never install anything.

The Clevid Desktop Client is a lightweight desktop application the candidate runs only during the exam. It checks display readiness, blocks floating overlays and unauthorized window switching, refuses to run inside a virtual machine or remote desktop session, and records synchronized screen, webcam, and audio streams. When the candidate submits, it displays a clear session receipt and closes completely.

Between them sits a server that owns the timer, grading, and recorded evidence. Neither the browser nor the candidate’s computer clock is trusted with scoring or timing.

Three rules we will not break

1. No flag fails a candidate automatically. Every integrity signal goes to a reviewer on the hiring team with the matching video clip attached. They can mark the session verified, request a retake, or confirm a violation. Every action is logged.

2. We state what software can and cannot do. Our documentation explains what the desktop app checks and the physical-room limits that require human review. We never install permanent system drivers or background services that run after the test ends.

3. Candidates are users too. Candidates get a clear preflight check before the timer starts, a plain-language session receipt afterwards, an appeal path, and free practice courses on the exact subjects companies test.

What ships today

Coding challenges, multiple-choice, short-answer, and typing assessments. Question banks organized by skill and difficulty tier so every candidate receives an equivalent, unique paper. Bulk CSV invitations, ranked cohort tables, expandable evidence dossiers, and team audit logs.

Try it

The Free plan gives you ten proctored attempts a month with every capability above included. Invite yourself first—run a test in the desktop client, inspect the evidence timeline, and see how it works end to end. Start free.