MedAI Guardian is an offline, locally running healthcare assistant powered by gpt-oss reasoning models. It provides basic medical guidance, triage, and emergency first-aid instructions without requiring internet access — making it life-saving for rural areas, disaster zones, and underserved regions.
Key Features:
🧠 Offline Reasoning: Uses gpt-oss models locally to answer healthcare queries without internet.
💊 Symptom Checker: Patients describe symptoms, and the assistant suggests possible causes + urgency level.
🚑 Emergency Mode: Guides step-by-step for CPR, burns, bleeding, and other first-aid emergencies.
📚 Localized Knowledge: Fine-tuned with WHO health manuals + local healthcare guidelines.
🔊 Voice Interaction: Patients can ask questions via speech, making it useful for illiterate or visually impaired users.
🌍 Multi-language Support: Runs in English, Hindi, and other regional languages.
🔋 Runs on Low Power Devices: Raspberry Pi, Android phones, or low-cost laptops.
This project tackles the global healthcare gap by ensuring that reliable medical knowledge is available to anyone, anywhere, anytime — no internet required.
Built With
- c++-(for-lightweight-hardware-integration)-frameworks:-fastapi
- javascript-(for-ui)
- languages:-python
- pytorch-(for-model-execution)-databases:-sqlite-(local-patient-record-storage)-cloud/tools-(optional):-hugging-face-for-initial-fine-tuning
- react-native-(for-mobile-app)
- speech-to-text
- then-fully-offline-hardware-(optional-for-demo):-raspberry-pi-5-+-microphone-+-small-touchscreen-apis-(local):-text-to-speech-(coqui-tts)
Log in or sign up for Devpost to join the conversation.