mtl-autonote-web

AutoNote Web SDK

AutoNote Web SDK is a JavaScript library that enables real-time transcription and automated medical report generation from doctor-patient conversations. It uses Voice Activity Detection (VAD) to record and process speech, then generates structured medical reports.

Installation

npm install mtl-autonote-web

Features

• Real-time voice activity detection
• Automatic transcription of doctor-patient conversations
• Generation of structured medical anamnesis reports
• Configurable speech detection parameters

Usage

Basic Implementation

import AutoNote from 'mtl-autonote-web';

const autoNoteInstance = new AutoNote({
  baseUrl: 'YOUR_API_BASE_URL',
  email: 'YOUR_EMAIL',
  password: 'YOUR_PASSWORD'
});

// Set up callback for receiving the report
autoNoteInstance.getReportResult((result) => {
  console.log('Transcription:', result.transcription);
  console.log('Report:', result.report);
});

// Start recording
autoNoteInstance.startRecording();

// Later, stop recording and generate report
autoNoteInstance.stopRecording();

API Reference

AutoNote(options)

Initializes the AutoNote SDK and returns functions to control recording and report generation.

Parameters

Parameter	Type	Required	Default	Description
baseUrl	string	Yes	-	Base URL for the API endpoint
email	string	Yes	-	Authentication email
password	string	Yes	-	Authentication password
positiveSpeechThreshold	number	No	0.45	Threshold for detecting speech (0.0-1.0)
negativeSpeechThreshold	number	No	0.25	Threshold for detecting non-speech (0.0-1.0)
redemptionFrames	number	No	20	Number of frames to wait before ending speech detection
frameSamples	number	No	1536	Number of audio samples per frame
preSpeechPadFrames	number	No	2	Number of frames to include before speech is detected
minSpeechFrames	number	No	2	Minimum number of frames required to register as speech

Returns

An object containing the following methods:

Method	Description
startRecording()	Begins capturing audio through the microphone
stopRecording()	Stops recording and processes the captured audio
getReportResult(callback)	Sets a callback to receive the transcription and report results