Acceptable use

Draft. Needs review by a qualified lawyer before it ships — this is a starting point that says what the product is for and what it must not be used for, not a finished legal instrument.

What this product is for

Overpane reads what is on your screen and answers questions about it, in a panel that stays out of a shared screen or a screenshot. It is built for:

• Live meetings — recalling a figure, drafting an answer, summarising a document someone just put on screen.
• Coding and debugging — explaining an error, proposing an approach, reviewing code you are looking at.
• Design work — talking through an architecture or a class model.
• Practice and preparation — rehearsing on your own, before the real thing.

What it must not be used for

The panel is excluded from screen capture so that your own private notes stay private. That is a confidentiality feature, in the same category as a password manager hiding itself from a screen recording. It is not a licence to conceal assistance from someone who is entitled to know about it.

You must not use this product:

• In any proctored or invigilated exam, certification test, or academic assessment.
• In a hiring assessment, technical interview, or take-home evaluation where outside assistance is not permitted by the organisation running it.
• To circumvent, disable, or defeat any monitoring, integrity, or anti-cheating system you have agreed to run.
• To capture or process another person's screen, or a screen you are not authorised to read.
• To process data you are contractually or legally barred from sending to a third-party AI provider — check this before pointing it at anything under an NDA, and note that the bundled-inference tier sends screen text off the device while the bring-your-own-agent tier's traffic goes wherever your own provider sends it.

Where the rules of an exam, interview, or workplace conflict with anything above, their rules win. If you are unsure whether assistance is permitted, ask the organisation running the process — not us.

What we can and cannot enforce

Stated plainly, because a policy that implies more than it can do is worth nothing.

The application runs entirely on your machine. In the bring-your-own-agent tier it does not report to us at all. We cannot see what is on your screen, cannot tell what application you are using it over, and cannot detect a violation of this policy on our own. We are not going to pretend otherwise.

What we can do, and will:

• Not market, describe, or demonstrate this product as a way to pass interviews, exams, or assessments — including in advertising, documentation, demos, and anything we say about it publicly.
• Act on credible reports. An account used in violation of this policy may be suspended or terminated, and licences may be revoked, without refund.
• Cooperate with lawful requests from an institution or employer where we are legally obliged to.

Consequences

Violating this policy may end your licence. It may also expose you to consequences from your employer, your institution, or the organisation running an assessment — including a withdrawn offer, a failed result, disciplinary action, or a claim against you. Those consequences are yours, not ours, and this policy does not shield you from them.

Reporting

If you believe this product is being used in breach of this policy, contact [email protected]. Include what you observed and when. We will look into it.

Related: the technical properties this application holds are documented with the application itself, and the privacy boundary this policy sets the limits of is described in the documentation.