Signing in
It says sign-in is unavailable. That is a configuration problem on our side, not your account. Try again shortly, or write to us.
No confirmation email after signing up. Check the spam folder. Sign-in works only after the confirmation link is opened.
Live assistant
The interviewer's audio is not coming through. Check against the capture route you picked. Speakers: make sure you are not wearing headphones. Browser tab: make sure "Also share tab audio" was ticked when you started. Desktop app: make sure "Entire screen" was chosen and "Also share system audio" ticked. On macOS with Chrome below 141, install a virtual audio device first.
It says this browser cannot listen. Use Chrome or Edge.
Audio stopped mid-interview. Do not change audio devices during an interview; plugging or unplugging headphones cuts the capture. End and restart the live assistant.
When using speakers, my reply makes the previous prompt disappear or regenerate. The current version locks the prompt after generation and logs your speech as you. If you still see the old behaviour, reload the page and restart the live assistant. Choose “Continue to next question” after each answer; this explicit boundary prevents a mixed microphone feed from treating your reply as a question.
The voice trigger did not work. Confirm the capture route is “Computer speakers”, the live panel is set to “Voice trigger”, and the question is visible in the conversation transcript. Say “Let me think” clearly after the interviewer finishes. You can always choose “Generate suggested answer” or use the shortcut; both use the same pending question.
Automatic prompting is too early or too late. Auto is available only for a shared tab or system audio. Pick quick (1.0s), standard (1.6s) or relaxed (2.2s) in the live panel. Use relaxed when the interviewer pauses while thinking.
The question was transcribed incorrectly. Choose “Edit question” on the prompt card, edit it and select “Save and regenerate”. Earlier questions stay in history; only this explicit action stops the incorrect in-flight request.
The second screen shows nothing on my phone. The second screen has to be another tab or window of the same browser. A phone is a separate device and does not receive the session.
Extension
It says the extension is missing, but I installed it. Reload the Greenroom page after installing. Until then the page has no channel to the extension.
It says access has not been granted. Choose "Grant access" in the interface; the extension opens its own page for you to confirm. This cannot be done from inside Greenroom.
A form says this site is not supported. That hiring system has no adapter yet. Use "Open the application page" and fill it in yourself; the rest of the material is still there in the opportunity.
Company status pages all land on a login screen. Sign in to that hiring system in your browser first — the extension uses the session you already have.
Material and generation
It says no resume has been imported. Nearly everything generated is built on your resume. Import one in Studio first.
What was generated does not match my experience. Check the material it was built from. The script uses only facts present in your material, so a mismatch usually means the original is unclear. Open any figure's source mark to check where it came from.
The last answer stopped halfway. Closing the page while text is being generated cuts it off. Ask again; the interrupted call is not charged.
Contact
privacy@ungetsu.net. When reporting a problem, say which screen you were on, what you did, and what the interface said. Please do not attach your resume or interview notes — they are not needed to find the fault.