Feature guide

Real-Time Word Highlighting

Advanced speech recognition technology that brings instant, word-by-word visual feedback to pronunciation practice. Learn how it works, try the demo, and understand browser and language compatibility.

Status
Not supported
Browser
Unknown Unknown

Feature status

Fallback mode

Language support

Full support

Auto-stop delay

0.5s

Device detected

Desktop

What is Real-Time Word Highlighting?
A modern feedback experience that mirrors how classrooms and coaches give corrections, but in software.

How it works

  • Uses the Web Speech API with interim results enabled for live transcripts.
  • Processes speech in real time and compares the transcript to the target sentence.
  • Aligns each spoken token to the sentence and updates highlights instantly.
  • Applies visual states to show progress, pacing, and any missed words.

Key benefits

  • Immediate feedback on pronunciation accuracy and timing.
  • Higher engagement during practice sessions with visual progress cues.
  • Automatic recording stop when the sentence is complete.
  • Lower cognitive load thanks to clear instructional visuals.
Visual states & color coding
Each highlight state is designed to read like a professional learning report while remaining accessible.

Current word

The word you are actively speaking, highlighted with a subtle pulse for clarity.

Spoken words

Words that have already been confirmed, helping you track overall progress.

Next word

A preview of the upcoming word to keep the pace natural and confident.

Missed words

Words that were skipped before the session ended, surfaced for quick review.

Auto-stop workflow
Intelligent recording termination keeps the experience natural while avoiding accidental cut-offs.

Primary workflow

  1. 1

    Word detection

    Each spoken word is matched in real time against the practice sentence.

  2. 2

    Completion check

    The system looks for near-completion before preparing to stop.

  3. 3

    Delay timer

    A configurable pause confirms that you really finished speaking.

  4. 4

    Auto-stop

    Recording ends automatically once the delay expires without new words.

Fallback to silence detection

When real-time highlighting is unavailable, silence detection preserves auto-stop accuracy without additional configuration.

  1. 1

    Audio monitoring

    Continuously inspects audio levels to understand when speech is present.

  2. 2

    Silence detection

    Identifies silence windows when voice activity falls below the threshold.

  3. 3

    Delay timer

    Uses the same delay value to avoid cutting off lingering syllables.

  4. 4

    Auto-stop

    Stops recording automatically when silence is confirmed.

Real-time word highlighting guide · Last updated December 2024

Estimated reading time ~8 minutes · Feature status: Fallback mode