The first rubric-based AI evaluator for Flowcharts, Algorithms, and Pseudocode. Transform grading from a manual bottleneck into an instant, insightful learning opportunity.
Traditional regex-based graders fail at logic. EVAPRO uses hybrid embeddings (LLaMA + MiniLM) to understand the *intent* behind the student's solution.
Don't limit students to text boxes. We accept handwritten Flowcharts (OCR), Word docs, PDFs, and raw text.
Our engine captures semantic similarity. A correct algorithm written in different words is still correct.
Pseudocode has no strict syntax. We evaluate the logic flow independent of the specific notation style used.
Sync grades directly to Canvas, Blackboard, or Moodle. No manual CSV exports needed.
Four simple steps to automate your entire grading workflow, from rubric creation to final feedback generation.
Try the interactive demoCreate dynamic rubrics with weighted criteria. Define what constitutes a 'logical error' vs a 'syntax error' for your specific assignment.
Students upload their work. Whether it's a photo of a whiteboard flowchart or a PDF of an algorithm, we ingest it all.
Our FastAPI backend splits the document, OCRs visual elements, and runs semantic comparison against your rubric using Vector Embeddings.
Students receive detailed, constructive feedback instantly. Instructors get a dashboard of performance metrics.
Join forward-thinking institutions using EVAPRO to provide faster, fairer, and more consistent feedback.