FAQ

Is Invisible Interview really undetectable?

Yes — we’re built for stealth via native, focus‑safe behavior and stream masking.

• Focus‑safe surfaces that don’t trigger active‑window switches
• No browser DOM overlays or content scripts to be detected
• Stream masking in common conferencing/capture setups

Electron vs native: what can’t Electron reliably hide?

Electron apps are great for many products, but the Chromium windowing stack is still a standard top‑level window from the OS perspective. That makes it more likely to be captured, focused, or flagged by platform heuristics. Native approaches can leverage non‑activating presentation, system composition layers, and window attributes that aren’t available (or are brittle) in Electron.

• Native window attributes designed to avoid focus stealing
• OS composition behaviors that keep guidance off shared streams
• No injected DOM/UI into proctored pages

Will it steal focus or create suspicious popups?

No. Guidance appears in non‑activating, focus‑safe views. Your coding environment remains the active application. That keeps cursor behavior and typing natural during interviews.

• No unexpected app focus changes
• No alt‑tab clutter or intrusive overlays
• Consistent typing feel during OAs and live rounds

Will Zoom, Google Meet, or Teams capture it?

In common setups, our guidance is designed not to appear in standard window/tab capture paths. For edge cases, we document safe capture settings and offer quick self‑checks so you can verify your environment before an interview.

Tip: Use Zoom 6.16 or older, or enable Advanced capture with window filtering. See our basic checks: /detection_tool.html

Does it type for me? How are keystrokes handled?

We prioritize human‑like workflows. By default, we avoid synthetic typing bursts and clipboard patterns that look automated. You stay in control: read the reasoning, copy snippets, and paste when ready. That reduces the risk patterns typically associated with automation tools.

How do you extract structure from screenshots?

Our on‑device ingest combines OCR with a constraint‑aware parser to identify inputs, outputs, edge cases, and test scaffolding. You get reasoning tailored to the specific prompt — not generic boilerplate.

Is it safe with CodeSignal/HackerRank proctoring?

Yes. We don’t inject scripts, modify the DOM, or overlay browser UI on top of proctored pages. Guidance lives outside the page’s process and is designed to remain invisible to typical page‑level checks.

Why is this safer than browser extensions?

Extensions interact directly with the page. That leaves fingerprints: content scripts, DOM mutations, or network hooks. Invisible Interview avoids that surface entirely and keeps assistance out of the browser’s security context.

Which platforms and versions are supported?

We support major interview platforms (HackerRank, CodeSignal, CoderPad) and conferencing tools (Zoom, Meet, Teams). For Zoom, use 6.16 or older — or enable Advanced capture with window filtering. See the Compatibility section for current status.

Is there a free trial?

Yes. Try the free tier — limited models and partial solutions — then upgrade when you’re ready.

Is usage unlimited on Pro?

We show “Unlimited” in the UI for a clean experience, but usage is still governed by a fair use policy to keep the service healthy for everyone. Pro is capped at 1,000 messages/day and 12 hours/day of audio.

What about privacy?

Ingest and reasoning run locally with minimal telemetry. We don’t inject into interview pages, and we encourage you to run our basic checks to verify your capture setup. See our Privacy Policy: /docs/privacy