The product, in full
Every capability, including the ones you could only see after logging in.
Proctor is a native agent, a live session view, an evidence store a person reviews, and the platform work around all three. Here is the whole list — with what it deliberately will not do at the end of it.
Already a customer? Sign in
- Capabilities
- 22
- Desktop platforms
- 3
- Risk scores produced
- 0
The candidate reads the scopes, then agrees to them
The consent screen names the exact scopes, and the server refuses any capture or upload that arrives without a matching consent record. Once the session starts, an indicator stays visible until it ends.
- Screen, camera and microphone are separate scopes, agreed to individually.
- The ingest API rejects a write with no consent record — the rule is enforced on the server, not in the UI.
- No hidden mode exists to enable, for administrators or anyone else.
A record a reviewer can read — and verify
Timeline, media and transcript sit in one view, and every action is written to an append-only log where each row carries the digest of the one before it. A row that was removed or edited does not go unnoticed.
- Observations are described in words, never converted into a number.
- Evidence is encrypted at rest and served through short-lived, access-checked reads.
- Retention windows expire it on schedule; a legal hold suspends expiry for named sessions only.
Provision a session from your own systems
The API is the product's front door, not an afterthought bolted to a dashboard. Create sessions, fetch evidence, set retention and subscribe to events — from your ATS, your internal tooling, or a script.
- A published OpenAPI document, with the Node SDK generated from it.
- Signed webhooks with a delivery log and replay for the attempts that failed.
- Scoped API keys per environment, so a test key cannot touch production evidence.
The same call the quickstart opens with. Full reference in the API docs.
Everything, grouped by where it lives
Anything not shipped yet is marked as such. Nothing on this list describes the product as detecting dishonesty, because it does not.
On the candidate's machine
A native app that captures only what was agreed to, and says so for the whole session.
Native desktop agents
Swift on macOS, WinUI 3 on Windows, GTK4 on Linux — each using its platform's real capture and permission APIs, and installed as a normal application.
Consent-gated capture
No capture, event or upload happens until the server holds a consent record naming the exact scopes. Without one, the ingest API refuses the write.
Always-visible indicator
A banner in the app and an indicator in the menu bar or tray, for the whole session. There is no hidden mode to enable — not even for administrators.
Scoped to what was agreed
Screen, camera and microphone are separate scopes. What the candidate did not agree to never crosses the network — no keystrokes, no window titles, no software inventory.
Buffers through a dropout
If the network drops, the agent buffers locally and resumes when it comes back. A bad connection costs you a few seconds, not the session.
Paired mobile side view
An optional second angle from the candidate's phone, paired for the duration of the session and dropped when it ends. The phone apps are not yet listed in the app stores.
While the session runs
Watch it live, with the people who should be there and nobody who should not.
Live session view
Watch a running session from the browser — media, participants and the event stream as it is written.
Roles and presence
See who is in the session and in what role, including read-only observers who can watch but not interact.
Second-angle review
The paired phone's view sits beside the main capture, so the desk and the screen are read together rather than in two separate tools.
Reconnect-safe transport
Viewers and agents reconnect without restarting the session or losing the events written while they were away.
Evidence a person reviews
Everything the session produced, kept tamper-evident and readable in one place.
Session viewer
Timeline, media and transcript in one view, so a reviewer reads a moment and its context together instead of correlating three exports.
Append-only audit log
Every action is written to a hash-chained log where each row carries the digest of the one before it, so a removed or edited row is detectable rather than silent.
Evidence encrypted at rest
Session objects are encrypted in storage and served through short-lived, access-checked reads — never a public link.
Retention windows and legal hold
Evidence expires on the window you set, per source. A legal hold suspends expiry for the sessions it names, and only those.
Compliance pack
Export what was consented to, what was captured, where it went and when it expires — as one document a works council or auditor can actually read.
In your stack
The API, identity, roles and metering that make this an internal system rather than a tool someone bought.
REST API and Node SDK
Provision sessions, fetch evidence and manage retention from your own systems. The OpenAPI document is published, and the SDK is generated from it rather than hand-written.
Signed webhooks
Subscribe to session events with a signed secret, inspect every delivery attempt, and replay the ones that failed.
ATS and meeting integrations
Connectors for the systems you already schedule in, so a session is provisioned where the interview is booked. Each connection is set up and tested with you before live candidates go through it.
SSO with SAML, OIDC and LDAP
Enforce your identity provider across the organisation, map groups to roles, and provision users with SCIM.
Custom roles and permissions
Start from the built-in roles or define your own, down to the individual permission. A reviewer's access is not an administrator's.
White-label branding
Put your own name, logo and palette on what the candidate sees, so the session does not look like it arrived from a third party.
Usage metering
Sessions, storage and seats counted and reconciled, visible to your administrators as they accrue.
What it will not do
- It produces no score, no risk rating and no pass/fail. Observations are described; a named person draws the conclusion.
- It does not log keystrokes, read window titles, or inventory installed software — in any mode, ever.
- There is no hidden or silent mode. If it is capturing, the candidate can see that it is.
- It does not claim to prove identity or to detect dishonesty. It records what was observed, under consent, for a person to interpret.
If a capability is not on this page, we do not have it. If a limit is on this page, it is enforced in the product — not promised in a policy document.
See it against a real interview
Read the quickstart to see the whole loop, then ask for access — a person sets up your organization with you.
Already a customer? Sign in