How saidmansurhon/mock-cefr works
This is a CEFR speaking-test web app: a learner receives a multi-part test, prepares and records spoken answers, then sees an AI-generated level, explanation, and tip. The React app controls test and part progression; the Express API supplies test content, accepts audio, transcribes it, and evaluates the collected answers. MongoDB stores test definitions and GridFS images; Deepgram and OpenAI provide speech recognition and assessment. The source excerpts show these main flows; the frontend README is present but contains no architecture details, and other frontend modules were not sampled.
Main entry points: index.js Test API · App.js Session coordinator · SpeechTestLogic.js Recording flow · data_loader.js Content loader
Components 15
Test experience
- Session coordinator
speaking-test/frontend/src/App.js - Recording flow
speaking-test/frontend/src/SpeechTestLogic.js - Test interface
speaking-test/frontend/src/SpeechTestUI.js
Test services
- Test API
speaking-test/backend/index.js - Session answers
speaking-test/backend/index.js - Image delivery
speaking-test/backend/index.js
Test content
- Test definitions
speaking-test/backend/index.js - Content loader
speaking-test/backend/data_loader.js - Question data
speaking-test/backend/questions.json - Image assets
speaking-test/backend/images
Speech assessment
- Final evaluator
speaking-test/backend/index.js - Deepgram
- OpenAI
Other components
- Learner
- Microphone
Connections 25
- Learner → Session coordinator: uses
speaking-test/frontend/src/App.js - Session coordinator → Test API: starts test
speaking-test/frontend/src/App.js - Test API → Test definitions: selects test
speaking-test/backend/index.js - Test API → Session answers: creates session
speaking-test/backend/index.js - Session coordinator → Recording flow: renders part
speaking-test/frontend/src/App.js - Recording flow → Test interface: renders controls
speaking-test/frontend/src/SpeechTestLogic.js - Learner → Test interface: answers prompts
speaking-test/frontend/src/SpeechTestUI.js - Test interface → Recording flow: starts recording
speaking-test/frontend/src/SpeechTestUI.js - Recording flow → Microphone: captures audio
speaking-test/frontend/src/SpeechTestLogic.js - Recording flow → Session coordinator: submits answer
speaking-test/frontend/src/SpeechTestLogic.js - Session coordinator → Test API: uploads audio
speaking-test/frontend/src/App.js - Test API → Deepgram: transcribes audio
speaking-test/backend/index.js - Test API → Session answers: stores transcript
speaking-test/backend/index.js - Test API → Final evaluator: requests assessment
speaking-test/backend/index.js - Final evaluator → Session answers: reads answers
speaking-test/backend/index.js - Final evaluator → Test definitions: reads questions
speaking-test/backend/index.js - Final evaluator → OpenAI: requests rating
speaking-test/backend/index.js - Test API → Session coordinator: returns result
speaking-test/frontend/src/App.js - Content loader → Question data: reads data
speaking-test/backend/data_loader.js - Content loader → Image assets: reads assets
speaking-test/backend/data_loader.js - Content loader → Test definitions: loads definitions
speaking-test/backend/data_loader.js - Content loader → Image delivery: uploads images
speaking-test/backend/data_loader.js - Test API → Image delivery: serves images
speaking-test/backend/index.js - Recording flow → Image delivery: builds image URLs
speaking-test/frontend/src/SpeechTestLogic.js - Session coordinator → Test API: fetches final
speaking-test/frontend/src/App.js