Voice-Enabled Medical Tutor Source
Budget: ₹75,000 – ₹150,000 INR
I need complete, well-commented code for an AI tutor that learners can talk to rather than click through. The system must do two things out of the box: run medical-knowledge quizzes and guide users through interactive case studies. Content will initially cover general medicine, paediatrics, and surgery, so the knowledge base has to be modular and easy for me to extend later.
The interface will be fully voice-assisted. When a learner asks for a quiz or a case, speech-to-text should capture the request, the engine selects the right question or scenario, then text-to-speech delivers feedback in real time. Open-source Python libraries such as Vosk, Whisper, Google Speech API, pyttsx3, or gTTS are fine as long as they can be swapped out or configured in a single settings file. I plan to run everything locally, so please avoid heavyweight dependencies and stick to permissive licences.
Acceptance criteria
• One-command launch on Windows/macOS.
• Response latency under two seconds on a mid-range laptop.
• Quiz questions and case files editable in plain JSON or YAML.
• Clear inline comments plus a concise README covering setup, architecture, and how to add new material.
Deliverables
1. Complete, functioning source code.
2. Sample dataset: at least 30 quiz questions and 3 multi-step cases spread across the three subjects.
3. README with step-by-step installation and extension guide.
4. (Nice to have) short screencast showing voice interaction end-to-end.
The interface will be fully voice-assisted. When a learner asks for a quiz or a case, speech-to-text should capture the request, the engine selects the right question or scenario, then text-to-speech delivers feedback in real time. Open-source Python libraries such as Vosk, Whisper, Google Speech API, pyttsx3, or gTTS are fine as long as they can be swapped out or configured in a single settings file. I plan to run everything locally, so please avoid heavyweight dependencies and stick to permissive licences.
Acceptance criteria
• One-command launch on Windows/macOS.
• Response latency under two seconds on a mid-range laptop.
• Quiz questions and case files editable in plain JSON or YAML.
• Clear inline comments plus a concise README covering setup, architecture, and how to add new material.
Deliverables
1. Complete, functioning source code.
2. Sample dataset: at least 30 quiz questions and 3 multi-step cases spread across the three subjects.
3. README with step-by-step installation and extension guide.
4. (Nice to have) short screencast showing voice interaction end-to-end.