A desktop app that analyses video-call interviews live — identity, gaze, transcription — and produces an evidence-backed report afterwards.
Remote interview integrity checks are manual and subjective, and leave no evidence trail when something looks wrong.
- Tracks participants across frames using face embeddings and scores identification confidence live
- Flags suspicious gaze patterns with evidence screenshots
- Reads on-screen display names via OCR to corroborate identity
- Transcribes the recording afterwards and generates a report with findings
- Gives a six-tab review per session: report, replay, timeline, participants, transcripts and evidence
Split by capability
Electron owns capture, UI and local storage; a Python service owns every AI model. Each runtime does what it is good at, across one local boundary.
Evidence, not verdicts
Every score change is logged with its source and timestamp, so any conclusion traces back to the frame that produced it — which is what makes the output defensible.
Degrades cleanly
If one model fails to load, that feature switches off instead of taking down the pipeline.
If you need live video or audio analysed as it happens, with a record you can audit later.
Need something like this?
Tell us what you are trying to build. A few lines is enough to start, and you get a reply within one business day.