How amanpriyanshu/medtranslate-360 works
MedTranslate 360 is a Streamlit medical-note assistant. A clinician records a note; the app transcribes the audio, then on request translates the transcript and separately applies identifier-censoring prompts for a privacy-protected note. The README also describes automatic pause after two seconds of silence. The sampled app shows the orchestration and model calls; capabilities are represented only where documented or visible, and the graph does not imply a persistent clinical-record store or verified HIPAA compliance.
Main entry points: app.py Streamlit Interface · sample.wav Recorded Audio
Components 11
Clinician Interaction
- Streamlit Interface
app.py - Audio Capture
app.py - Note Results
app.py
Note Processing
- Recorded Audio
sample.wav - Speech Transcription
app.py - Medical Transcript
app.py - Note Translation
app.py - Identifier Censoring
app.py
AI Services
- Gemini Service
- Whisper Model
Other components
- Clinician
Connections 16
- Clinician → Streamlit Interface: uses
README.md - Streamlit Interface → Audio Capture: starts capture
app.py - Audio Capture → Recorded Audio: writes audio
app.py - Streamlit Interface → Speech Transcription: invokes
app.py - Speech Transcription → Recorded Audio: reads audio
app.py - Speech Transcription → Whisper Model: transcribes with
app.py - Speech Transcription → Medical Transcript: produces
app.py - Streamlit Interface → Note Translation: invokes on request
app.py - Note Translation → Gemini Service: generates translation
app.py - Streamlit Interface → Identifier Censoring: invokes on request
app.py - Identifier Censoring → Gemini Service: generates censored note
app.py - Note Translation → Medical Transcript: uses transcript
app.py - Identifier Censoring → Medical Transcript: uses transcript
app.py - Streamlit Interface → Note Results: displays
app.py - Note Translation → Note Results: supplies translation
app.py - Identifier Censoring → Note Results: supplies censored note
app.py