Hear a candidate answer. See what the scorecard makes of it.
Fifty practice roles, from frontend code to structural steel: software, systems and cloud, data and AI, embedded and hardware, mechanical and industrial engineering, sales. In each one, candidates of different quality answer the same questions: specific, vague, reading from a script, solid but thin, overclaiming. Play any answer, then read how Interview Signal scored it: the quotes it kept, the areas it left to probe, and where it thought the answer was not their own.
A full interview in the app, start to finish
Recorded on the backend practice role below. The questions go into the microphone, the answers play through the computer's sound the way a video call does, and nothing on screen is typed in: the kit, the audio check, follow-ups, the scorecard filling in, an answer check on a rehearsed answer, and the client submittal.
The video plays from YouTube in privacy-enhanced mode, and nothing loads from YouTube until you press play. Watch it on YouTube.
Loading the practice interviews…
Pick an answer.
Run it through the app, live
The answers above are what the app produced from the transcripts. To see it happen in real time — audio in, transcript, follow-ups, scorecard — play them to the app:
- Install Interview Signal and sign in.
- On the home screen, press Add practice roles. You get the two full interviews from this page, backend and sales, with the same questions and scorecard.
- Start an interview on a practice role, choose Video call, test your audio and pick a consent option.
- Read each question aloud yourself, then press Play the answer here. The app hears it the way it hears a candidate on a call — through your computer's sound — and scores it as it arrives.
Play the answers through speakers or headphones; either works on Windows. On a Mac the app needs Screen & System Audio Recording turned on to hear them. If you press Play the question as well, the app hears the question as candidate audio too, so reading it aloud is the realistic test.
What you are looking at
Only the areas the answer speaks to
An answer about a missed quarter says nothing about compensation, so that area stays blank instead of getting a guess.
Every quote is checked
A quote the scorer cites has to appear, word for word, in what the candidate said. One that does not is deleted, and a score with nothing behind it is capped at 5.
Polish is not the same as proof
An answer full of complete, confident general claims and no names, numbers or decisions scores low, however good it sounds.
The answer check is a prompt, not a verdict
A flag says an answer reads as generated, gives the reasons, and suggests a question someone who did the work could answer at once. It never changes a score and never goes in the submittal.