Simple, predictable pricing.

ParikshAI grades both handwritten scanned answer booklets and CBT responses at scale — against rubrics you define — assigning a confidence score to every answer. Low-confidence answers are automatically flagged and routed to human evaluators before grades are finalised. Result timelines drop from 2 months to under 2 weeks.

99.8% accuracy measured against human evaluator scores across 50,000+ reviewed booklets. No grade is published without passing through the Evaluator Portal. Every AI score carries a confidence percentage — scripts below your set threshold are mandatory-reviewed by a subject evaluator before results are locked.

End-to-end — AI grading, confidence flagging, and human review — under 2 weeks for 10,000 scripts. The AI engine runs at 2,400 scripts per hour. Manual grading of the same batch typically takes 2–3 weeks before human review even begins.

Your institution owns all data — scripts, rubrics, and grades. ParikshAI is a data processor: we never use your data for model training. Data is encrypted in transit (TLS 1.3) and at rest (AES-256), stored on Indian data centre infrastructure, and deleted on request. A signed Data Processing Agreement is provided to every partner.

Answers below your confidence threshold cannot be published until a human evaluator approves them. For student disputes, ParikshAI provides a full per-script audit trail: the original scan, the rubric applied, AI score with reasoning, and every human action — satisfying university grievance redressal requirements.

Yes. ParikshAI accepts JPEG, PNG, TIFF, and multi-page PDF — standard output from MFD and document scanners already used by most examination departments. No specialist hardware required. For CBT, a REST API connects to Moodle, Mettl, and custom exam platforms.

ParikshAI is built with mandatory human-in-the-loop review, full audit trails, and institution-controlled grade locks — ensuring no AI score is published unilaterally. We work with each university's examination controller to map our workflow to their ordinance requirements. Compliance documentation is available for your academic council.

No technical staff needed. The Rubrics Studio is a no-code interface for subject faculty — define criteria, allocate marks, optionally upload a model answer — typically 15–30 minutes per question paper. Our onboarding team runs a rubric-building workshop with your department heads before the first batch.