Observations
What the candidate's device reported during the session — focus leaving the interview application, the side view pausing. Recorded facts about the session, never a judgement about the person.
In the report: Integrity signals
Illustrative sample
This is the report Proctor produces for every session, rendered from invented data. Read it the way a reviewer on your team would: observations first, then the evidence behind them, then a person's conclusion.
Synthetic data. Every name, identifier, time, count and note on this page is invented for illustration. It describes no real candidate, interviewer or organization.
What the candidate's device reported during the session — focus leaving the interview application, the side view pausing. Recorded facts about the session, never a judgement about the person.
In the report: Integrity signals
What was captured under the candidate's consent and what was actually stored, by type — so a reviewer knows what there is to look at, and what was never switched on.
In the report: Session · Captured, and Evidence inventory
The only conclusions in the document. Each note is attributed to a named reviewer, timestamped, and cannot be edited or deleted afterwards.
In the report: Reviewer notes
Generated 3/10/2026, 4:20:00 PM
What a named reviewer concluded about this session's evidence, and against which facts. It is kept apart from the observations below, and it is never a hiring decision.
rrv_sample00000000000000Signed by Sample Reviewer (synthetic) (Reviewer) · 3/10/2026, 4:20:00 PM
Signed against coverage revision sample0000000000 (incomplete) · 11 observations · 39 evidence objects
What the session's devices and this product recorded. Nothing in this part is a conclusion about the candidate.
Created from Sample policy — Backend live coding, version 3
Sample rules: keep your editor and the meeting window on screen. The session runs about 60 minutes with one short break.
Sample statement: language documentation is allowed. AI chat assistants may be used for syntax questions if you say so aloud.
Written by your organization and shown as written. The product never treats any AI or tool use as misconduct on its own — a reviewer decides.
What this session was set up and consented to collect, against what actually arrived. Counts alone cannot tell a lane nobody asked for from one that failed.
Incomplete — this session's collection does not support a statement about what happened. Read each lane below.
Present across the whole session, with no interval missing beyond the tolerance for ordinary timing drift.
Set up: yes · consented: yes · expected every 180s · reported as a gap after 360s · uploads in flight: 0 · device permission: granted
Present across the whole session, with no interval missing beyond the tolerance for ordinary timing drift.
Set up: yes · consented: yes · expected every 600s · reported as a gap after 1200s · uploads in flight: 0 · device permission: granted
This session was never set up to collect it, so there is nothing missing.
Set up: no · consented: no · expected every n/a · reported as a gap after n/a · uploads in flight: 0 · device permission: n/a
Present, with intervals that hold nothing. The intervals are listed; they are not interpreted.
Set up: yes · consented: yes · expected every 300s · reported as a gap after 600s · uploads in flight: 0 · device permission: n/a
Related observations grouped by time, in the order they happened. A moment is one episode, not a finding, and nothing here is ranked.
Brief version 1 · revision sample0000000000 · worked out 3/10/2026, 4:20:00 PM
Observations, not conclusions — a signal records what the device reported, never a judgment about the candidate.
| Type | Objects | Size |
|---|---|---|
| Screen recordings | 6 | 412.0 MB |
| Screenshots | 18 | 9.0 MB |
| Audio | 6 | 38.0 MB |
| Transcript | 1 | 64.0 KB |
| Side-view placement photos | 1 | 1.2 MB |
| Side-view recordings | 5 | 96.0 MB |
| Integrity logs | 1 | 180.0 KB |
| Timeline bundles | 1 | 40.0 KB |
| Total | 39 | 556.5 MB |
Transcript segments: 214
The human judgment on this session. Notes are attributed and never edited or deleted.
Sample Interviewer (synthetic) · 3/10/2026, 3:24:00 PM
Focus left the coding exercise twice, each time for under a minute. Both times the candidate said on the call that they were re-reading the brief we sent. Consistent with what I saw during the interview; nothing further to follow up on the session itself.
Sample Reviewer (synthetic) · 3/10/2026, 4:11:00 PM
Checked the side-view pause near the end: the candidate told us at the time that their phone battery had run low, and it resumed about three minutes later. No open questions about the session environment. The skill assessment is recorded separately in our interview scorecard.
Everything in the sample above is in the product now. The items marked planned are being built for hiring teams and are not in the report yet.
Tell us how your team interviews and a person will walk you through a real session and set up access.
Already a customer? Sign in