InterpreTutor is an AI-powered training and evaluation tool for spoken-language interpreting. It transcribes source and target audio, compares the interpreter’s output to a reference translation, and delivers structured, visual feedback.

The New Update

InterpreTutor just got a big upgrade focused on clarity, speed, and trust in results. Here’s what’s new.

• Visual dashboard, simplified
• Clean 1×3 layout: Overall score, performance gauge, and a concise score summary.
• No noisy charts or overlap. Just the essentials—readable at a glance.
• Smarter scoring from LLM outputs
• Robust parser extracts sub-criteria scores reliably:
  • Meaning & Intent
  • Deviations & Omissions
  • Overall Accuracy
  • Cohesion & Coherence
  • Terminology
  • Disfluency Markers
• AI-enhanced two-step evaluation with more robust models
• You choose the depth: new Step 5 adds a report style selector that sets how detailed the report will be.
  • → acceptably medium-length report, faster
  • → detailed report, a bit faster
  • → even more detailed report, but the slowest
• End-to-end progress tracking
• Waveform visualization for your interpretation.
• Quick metrics (duration, ZCR, RMS, spectral centroid, silence %).
• Reliability and UX polish
• Safer audio handling (temporary files for librosa).
• Cleaner labels, updated FAQ, and new branding.
• One-click export
• Download a formatted DOCX containing both evaluation parts.