Product documentation

Technical Documentation

Updated

A technical documentation hub for the Windows app, AI workflows, permissions, supported environments, and troubleshooting.

Documentation - FAQ - Security - Privacy - Responsible AI

Core trust links

Voice pipeline

A voice workflow typically captures audio from the user-selected Windows audio path, sends the audio or transcript through a speech-to-text step, and turns the resulting question into a concise answer outline. Manual input required: confirm exact desktop capture method and retention behavior.

Screenshot pipeline

A screenshot workflow captures a user-selected visible prompt, extracts or interprets the relevant text or visual context, and sends that context into the answer-generation workflow. Users should process only content they are allowed to share with an AI system.

Resume processing

Resume/profile context is used to make answer outlines less generic. Users should provide only accurate details and remove outdated or confidential information before using it in prompts.

Phone Mirror (second screen)

The Windows app can serve the current question and answer to a phone browser on the same local network. It opens a small local web page behind a random-key link shown as a QR code, streams only the changed part of the text while an answer is written, and reconnects open pages after Wi-Fi drops or an app restart. No cloud relay is involved, and the phone does no capture or answer generation.

Supported environments

  • Windows 10/11 desktop workflow.
  • Remote interview workflows such as Zoom, Google Meet, and Microsoft Teams.
  • Coding prompts visible in browsers, IDEs, documents, or assessment screens when users are allowed to process them.
  • Manual input for questions that cannot be captured reliably.

Troubleshooting

  • Restart the app before a high-stakes interview test.
  • Check microphone and output device permissions.
  • Confirm internet access and account login.
  • Try manual text input if voice capture is unclear.
  • Contact support with registered email and steps to reproduce.

Limitations

Some pipeline details, such as the exact desktop audio capture method and how long captured input is kept, still need confirmation and are marked on this page. Treat those parts as provisional.

All product limitations in the Trust Center

Frequently asked questions

What should I do if a spoken question is transcribed wrongly?

Type or paste the question instead, check that the right microphone and output device are selected, and retry in a quieter room.

Which content can the screenshot workflow read?

Visible prompts such as code, coding-platform questions, diagrams, tables, and logs. Small, blurred, or heavily compressed text can be misread, so check the extracted question before relying on the answer.

How is my resume used in answers?

Resume and profile context makes answer outlines less generic. Keep it accurate and remove outdated or confidential details before using it.

Which meeting tools does the workflow cover?

Remote interviews on Zoom, Google Meet, and Microsoft Teams on Windows 10 or 11, plus prompts visible in browsers, IDEs, documents, or assessment screens you are allowed to process.

What should I send support when something breaks?

Your registered email, Windows version, app version if available, the steps that cause the problem, and what you expected to happen.