P
ProveIQ
← Back to AI PillarsPillar 3

A score you can defend.

Cost per call:₹0 · In-house inference

Four signals combine into a single integrity score that runs alongside every evaluation. Low integrity = we tell the employer before they hire — never surprise rejections, never silent demotion.

API endpoint

POST /api/v1/integrity/check

Model routing

In-house specialized inference

Cost envelope

~₹0 per submission

Plagiarism search

Embeddings-based similarity against a corpus of past submissions, public GitHub repos in the same domain, and open-source snippets. High similarity flags for human review — not automatic rejection.

AI-content heuristics

Detects writing patterns and code patterns consistent with AI generation. Not deterministic (nothing is, for this) — produces a confidence score bundled with explanation cues.

Git history analysis

For code submissions, we examine commit cadence, diff granularity, and authorship signals. A repository that appears in its final state in a single commit five minutes before deadline looks very different from iterative work.

Webcam verification (opt-in)

For high-stakes evaluations, candidates can opt-in to webcam checkpoints during work sessions. Face-presence + screen-recording heuristics only. No biometric data stored beyond session. Granular DPDP consent required.