How krishnag180805-star/vocaliq---ai-english-evaluator works
VocalIQ is a Streamlit speaking-practice evaluator. Learners choose a structured or custom prompt, speak to a timer, record or upload audio, and receive a transcript plus diagnostic scores and feedback. The main runtime path runs through the app’s audio validation/conversion, speech recognition, and linguistic evaluation before rendering results. The README documents optional cloud recognition and AI feedback alongside offline evaluation. Source excerpts support the core orchestration and evaluation steps; some UI and integration links rely on README documentation because the relevant app and evaluator sections are incomplete.
Main entry points: app.py Streamlit app · evaluator.py Evaluation engine · stt_engine.py Speech recognition · audio_processor.py Audio validation
Components 21
Practice setup
- Streamlit app
app.py - Topic library
src/topics.py - Speaking timer
src/ui_components.py - Evaluation settings
src/config.py
Audio and recognition
- Mic or upload
app.py - Audio validation
src/audio_processor.py - Speech recognition
src/stt_engine.py - Local Whisper
- Google speech API
- Groq Whisper
- Audio recovery
src/error_handler.py
Assessment
- Evaluation engine
src/evaluator.py - Grammar analysis
src/evaluator.py - Vocabulary analysis
src/evaluator.py - Fluency analysis
src/evaluator.py - Pronunciation analysis
src/evaluator.py - Gemini AI
Results and guidance
- Evaluation report
src/evaluator.py - Result components
src/ui_components.py - Mic guidance
src/error_handler.py
Other components
- Learner
Connections 22
- Learner → Streamlit app: uses
README.md - Streamlit app → Topic library: selects prompts
README.md - Streamlit app → Speaking timer: shows timer
README.md - Streamlit app → Mic or upload: captures audio
README.md - Mic or upload → Audio validation: validates audio
README.md - Audio validation → Evaluation settings: reads limits
src/audio_processor.py - Streamlit app → Speech recognition: requests transcript
app.py - Speech recognition → Local Whisper: transcribes locally
src/stt_engine.py - Speech recognition → Google speech API: falls back
src/stt_engine.py - Speech recognition → Groq Whisper: optionally calls
src/stt_engine.py - Streamlit app → Evaluation engine: evaluates transcript
app.py - Evaluation engine → Evaluation settings: uses criteria
src/evaluator.py - Evaluation engine → Grammar analysis: analyzes grammar
README.md - Evaluation engine → Vocabulary analysis: analyzes vocabulary
README.md - Evaluation engine → Fluency analysis: analyzes fluency
README.md - Evaluation engine → Pronunciation analysis: analyzes clarity
README.md - Evaluation engine → Gemini AI: optionally enhances
README.md - Evaluation engine → Evaluation report: builds report
src/evaluator.py - Streamlit app → Result components: renders feedback
README.md - Evaluation report → Result components: displays scores
README.md - Streamlit app → Audio recovery: handles audio errors
app.py - Audio recovery → Mic guidance: shows guidance
src/error_handler.py