EkaScribe iOS SDK

A Swift package for voice-to-prescription functionality with audio recording and real-time transcription capabilities for medical consultation applications.

Overview

EkaScribe empowers healthcare applications with advanced voice recording and transcription capabilities. It provides a seamless integration for medical consultation workflows, enabling doctors to record patient interactions and automatically generate prescriptions through AI-powered voice analysis.

Getting Started

Install the package via Swift Package Manager
Configure authentication and doctor information at app launch
Initialize VoiceToRxViewModel in your view
Start recording with your desired templates and languages
Handle results when processing completes

The SDK requires iOS 17.0+ and provides a simple, straightforward API for voice-to-prescription functionality.

Key Features

🎙️ Voice Activity Detection (VAD) - Intelligent audio recording with automatic speech detection
🔄 Real-time Transcription - Live audio-to-text conversion during consultations
🏥 Medical Context Aware - Specialized for healthcare terminology and prescription generation
📊 Session Management - Complete recording session lifecycle management
☁️ Cloud Integration - Automatic audio upload and processing
📝 Template Support - Customizable output formats (SOAP notes, prescriptions, etc.)

Requirements

⚠️ Important: This SDK requires iOS 17.0 or later.

iOS: 17.0+
Swift: 5.9+
Xcode: 15.0+
Sdk Version: 1.3.1+

System Permissions

Add the following permissions to your app's Info.plist:

<key>NSMicrophoneUsageDescription</key>
<string>This app needs microphone access to record medical consultations</string>
<key>NSLocalNetworkUsageDescription</key>
<string>This app needs network access to upload and process audio recordings</string>

Installation

Swift Package Manager

Add EkaScribe to your project using Swift Package Manager:

In Xcode, select File → Add Package Dependencies

Enter the repository URL:

https://github.com/eka-care/EkaVoiceToRx.git

Or use SSH:

git@github.com:eka-care/EkaVoiceToRx.git

Choose the version or branch
Add to your target

Package.swift

dependencies: [
    .package(url: "https://github.com/eka-care/EkaVoiceToRx.git", from: "1.3.1")
]

Import Statement

After adding the package, import it in your Swift files:

import EkaVoiceToRx

Integration Guide

Follow these steps to integrate the SDK into your app:

Step 1: Configure Authentication and Doctor Information

Set up authentication tokens and doctor information when your app launches. This should be done once at app startup:

import SwiftUI
import EkaVoiceToRx

@main
struct YourApp: App {
    init() {
        configureEkaScribe()
    }
    
    var body: some Scene {
        WindowGroup {
            ContentView()
        }
    }
    
    private func configureEkaScribe() {
        // 1. Set authentication tokens (required)
        AuthTokenHolder.shared.authToken = "your_auth_token"
        AuthTokenHolder.shared.refreshToken = "your_refresh_token"
        AuthTokenHolder.shared.bid = "your_business_id"
        
        // 2. Configure doctor/user information (required)
        let config = V2RxInitConfigurations.shared
        config.clientId = "your_client_id"
        config.ownerName = "Dr. Smith"
        config.ownerOID = "doctor_oid_123"
        config.ownerUUID = "doctor_uuid_456"
    }
}

Important: Make sure to set authentication tokens and doctor information before making any SDK calls.

Step 2: Initialize the View Model

Create a VoiceToRxViewModel instance in your SwiftUI view or UIKit view controller:

SwiftUI:

@StateObject private var viewModel = VoiceToRxViewModel(
    voiceToRxInitConfig: .shared,
    voiceToRxDelegate: nil
)

UIKit:

private let viewModel = VoiceToRxViewModel(
    voiceToRxInitConfig: .shared,
    voiceToRxDelegate: nil
)

Step 3: Start Recording

Start a recording session with your desired configuration:

Parameters

The startRecording() method accepts the following parameters:

conversationType (VoiceConversationType): The type of conversation (see Conversation Types)
inputLanguage ([InputLanguageType]): Array of supported languages (see Input Languages)
templates ([OutputFormatTemplate]): Array of output format templates (see Template Management)
modelType (ModelType): The AI model to use for processing (see Model Types)

// Your implementation
private func startRecording() async {
    do {
        // 1. Fetch available templates (recommended)
        // See Template Management section for details on getTemplates API
        // VoiceToRxRepo.shared.getTemplates { result in ... }
        
        // 2. Configure templates
        let outputFormat: [OutputFormatTemplate] = [
            OutputFormatTemplate(
                templateID: "template-id-here",
                templateType: .defaultType,
                templateName: "SOAP Notes"
            )
        ]
        
        // 3. Set patient information (if applicable)
        V2RxInitConfigurations.shared.subOwnerOID = patientOID
        V2RxInitConfigurations.shared.subOwnerName = patientName
        
        // 4. Calling SDK function
        try await viewModel.startRecording(
            conversationType: VoiceConversationType.conversation,
            inputLanguage: [InputLanguageType.english, InputLanguageType.hindi],
            templates: outputFormat,
            modelType: ModelType.lite
        )
        
        // Successfully started recording
    } catch {
        // Handle errors
        await MainActor.run {
            handleRecordingError(error)
        }
    }
}

Note: It's recommended to use the getTemplates() API first to fetch available templates before creating the OutputFormatTemplate array. See the Template Management section for details.

Core Components

VoiceToRxViewModel

The central view model that manages the entire voice recording and processing workflow.

public class VoiceToRxViewModel: ObservableObject {
    @Published public var screenState: RecordConsultationState
    
    public init(
        voiceToRxInitConfig: V2RxInitConfigurations,
        voiceToRxDelegate: VoiceToRxDelegate?
    )
}

Recording States

The screenState property tracks the current state of the recording session:

public enum RecordConsultationState {
    case retry                          // Ready to retry after error
    case startRecording                 // Initial state, ready to start
    case listening(conversationType: VoiceConversationType) // Currently recording
    case paused                         // Recording paused
    case processing                     // Processing audio/generating prescription
    case resultDisplay(success: Bool, value: String?)  // Results available
    case deletedRecording              // Recording deleted
}

Conversation Types

public enum VoiceConversationType: String, CaseIterable {
    case conversation = "consultation"  // Doctor-patient conversation
    case dictation                     // Direct prescription dictation
}

Model Types

public enum ModelType {
    case pro      // High accuracy model
    case lite     // Faster, lighter model
}

Input Languages

public enum InputLanguageType {
    case english
    case hindi
    case tamil
    case telugu
    case kannada
    case malayalam
    case bengali
    case gujarati
    case marathi
    case punjabi
    case urdu
    case odia
    case assamese
}

Configuration

App-Level Setup

Configure the SDK when your app launches. This should be done once at app startup:

import SwiftUI
import EkaVoiceToRx

@main
struct YourApp: App {
    init() {
        configureEkaScribe()
    }
    
    var body: some Scene {
        WindowGroup {
            ContentView()
        }
    }
    
    private func configureEkaScribe() {
        // 1. Set authentication tokens (required)
        AuthTokenHolder.shared.authToken = KeychainHelper.fetchAuthToken()
        AuthTokenHolder.shared.refreshToken = KeychainHelper.fetchRefreshToken()
        AuthTokenHolder.shared.bid = JWTDecoder.shared.businessId
        
        // 2. Configure doctor/user information (required)
        let config = V2RxInitConfigurations.shared
        config.clientId = "your_client_id"
        config.ownerName = "Dr. Smith"
        config.ownerOID = "doctor_oid_123"
        config.ownerUUID = "doctor_uuid_456"
        
        // 3. Set up data container (optional - only if using SwiftData)
        // config.modelContainer = SwiftDataRepoContext.modelContext.container
    }
}

Important: Make sure to set authentication tokens and doctor information before making any SDK calls.

Session-Specific Configuration

Before starting a recording session, configure patient-specific information. This should be done each time you start a new recording:

// Set patient information for the current session (required before recording)
V2RxInitConfigurations.shared.subOwnerOID = selectedPatientOid
V2RxInitConfigurations.shared.subOwnerName = patientName

// Optional: Set appointment ID for context
V2RxInitConfigurations.shared.appointmentID = appointmentID

Note: Patient information can be set to empty strings if not applicable:

V2RxInitConfigurations.shared.subOwnerOID = ""
V2RxInitConfigurations.shared.subOwnerName = ""

Configuration Properties

public class V2RxInitConfigurations {
    public static let shared: V2RxInitConfigurations
    
    // Doctor/User Information
    public var ownerName: String?          // Doctor name
    public var ownerOID: String?            // Doctor OID
    public var ownerUUID: String?            // Doctor UUID
    public var clientId: String?             // Client identifier
    
    // Patient Information (set per session)
    public var subOwnerOID: String?         // Patient OID
    public var subOwnerName: String?        // Patient name
    
    // Optional Context
    public var appointmentID: String?       // Appointment identifier
    
    // Data Storage
    public var modelContainer: ModelContainer?  // SwiftData container (optional)
}

Authentication

Set authentication tokens before making any SDK calls:

AuthTokenHolder.shared.authToken = "your_auth_token"
AuthTokenHolder.shared.refreshToken = "your_refresh_token"
AuthTokenHolder.shared.bid = "your_business_id"

Recording Management

Starting Recording

Parameters

The startRecording() method accepts the following parameters:

conversationType (VoiceConversationType): The type of conversation (see Conversation Types)
inputLanguage ([InputLanguageType]): Array of supported languages (see Input Languages)
templates ([OutputFormatTemplate]): Array of output format templates (see Template Management)
modelType (ModelType): The AI model to use for processing (see Model Types)

// Start recording
try await viewModel.startRecording(
conversationType: VoiceConversationType.conversation,
inputLanguage: [InputLanguageType.english, InputLanguageType.hindi],
templates: templates,
modelType: ModelType.lite
 )

Note: It's recommended to use the getTemplates() API first to fetch available templates before creating the OutputFormatTemplate array. See the Template Management section for details.

Pausing and Resuming

// Pause recording
viewModel.pauseRecording()

// Resume recording
do {
    try viewModel.resumeRecording()
} catch {
    // Handle resume error
    print("Failed to resume: \(error.localizedDescription)")
}

Stopping Recording

// Stop recording and process
Task {
    do {
        try await viewModel.stopRecording()
        // The viewModel will transition to .processing, then .resultDisplay
    } catch {
        // Handle error
        print("Failed to stop recording: \(error.localizedDescription)")
    }
}

Handling Results

Monitor the screenState for result availability. Results are automatically provided when processing completes:

case .resultDisplay(success: let success, value: let value):
    if success {
        // Get the result text (base64 encoded)
        let base64EncodedText = value ?? ""
        
        // Decode base64 to get the actual result
        guard let decodedData = Data(base64Encoded: base64EncodedText),
              let resultText = String(data: decodedData, encoding: .utf8) else {
            showError("Failed to decode result")
            return
        }
        
        // Get the session ID for future reference
        let sessionID = viewModel.sessionID?.uuidString ?? ""
        // Display results to user
        showResults(resultText, sessionID: sessionID)
    } else {
        // Handle error
        showError("Failed to process recording")
    }

Note: The value parameter contains base64-encoded formatted output based on your selected template. You need to decode it before displaying to users. For more detailed results including multiple template outputs, use VoiceToRxRepo.shared.fetchResultStatusResponse().

Template Management

Fetching Available Templates

Use VoiceToRxRepo to fetch available templates:

import EkaVoiceToRx

VoiceToRxRepo.shared.getTemplates { result in
    switch result {
    case .success(let response):
        let templates = response.items
        // Use templates
        break
    case .failure(let error):
        print("Failed to fetch templates: \(error.localizedDescription)")
        break
    }
}

Updating Favorite Templates

Update user's favorite templates configuration:

let templateIDs = ["template-id-1", "template-id-2"]

VoiceToRxRepo.shared.updateConfig(
    templates: templateIDs
) { result in
    switch result {
    case .success:
        print("Templates updated successfully")
    case .failure(let error):
        print("Failed to update: \(error.localizedDescription)")
    }
}

Using Templates in Recording

Pass templates when starting a recording:

// Your implementation
private func startRecordingWithTemplates() async {
    do {
        // 1. Fetch available templates first (recommended)
        // VoiceToRxRepo.shared.getTemplates { result in ... }
        
        // 2. Configure templates
        let templates: [OutputFormatTemplate] = [
            OutputFormatTemplate(
                templateID: "template-id",
                templateType: .defaultType,
                templateName: "Template Name"
            )
        ]
        
        // 3. Calling SDK function
        try await viewModel.startRecording(
            conversationType: VoiceConversationType.conversation,
            inputLanguage: [InputLanguageType.english],
            templates: templates,
            modelType: ModelType.pro
        )
    } catch {
        // Handle error
        await MainActor.run {
            handleError(error)
        }
    }
}

Session History

Fetching Session History

Retrieve past recording sessions:

VoiceToRxRepo.shared.getEkaScribeHistory { result in
    switch result {
    case .success(let response):
        let sessions = response.data
        // Display sessions
        break
    case .failure(let error):
        print("Failed to fetch history: \(error.localizedDescription)")
        break
    }
}

Result Management

Fetching Full Result Response

Get the complete result response with selected template outputs:

VoiceToRxRepo.shared.fetchResultStatusResponse(sessionID: sessionID) { result in
    switch result {
    case .success(let response):
        // Note: All values in the response are base64 encoded and need to be decoded before display
        // response.data.templateResults.custom - custom template results (base64 encoded)
        // response.data.templateResults.transcript - transcript results (base64 encoded)
        // response.data.output - general output (base64 encoded)
        // response.data.audioMatrix - audio quality metrics
        
        // Example: Decoding a result value
        if let encodedValue = response.data?.templateResults?.custom?.first?.value,
           let decodedData = Data(base64Encoded: encodedValue),
           let decodedText = String(data: decodedData, encoding: .utf8) {
            // Use decodedText for display
            print("Decoded result: \(decodedText)")
        }
        break
    case .failure(let error):
        print("Failed to fetch response: \(error.localizedDescription)")
        break
    }
}

Note: The value parameter returned by fetchResultStatus is base64 encoded and must be decoded before displaying to users.

Switching Templates

Switch the output format for an existing session:

VoiceToRxRepo.shared.switchTemplate(
    templateID: newTemplateID,
    sessionID: sessionID
) { result in
    switch result {
    case .success(let response):
        // Updated response with new template output
        break
    case .failure(let error):
        print("Failed to switch template: \(error.localizedDescription)")
        break
    }
}

Updating Results

Update edited content back to the server:

let request = UpdateResultRequest(
    // Configure request with updated content
)

VoiceToRxRepo.shared.updateResult(
    sessionID: sessionID,
    request: request
) { result, errorStatus in
    switch result {
    case .success:
        print("Result updated successfully")
    case .failure(let error):
        print("Failed to update: \(error.localizedDescription)")
    }
}

Error Handling

Common Errors

The SDK provides specific error types:

public enum EkaScribeError: Error {
    case freeSessionLimitReached
    case microphonePermissionDenied
    case microphoneIsInUse
    case vadDetectorFailed
    case noSessionId
    case audioSessionSetupFailed
}

Error Handling Example

Here's a complete example of handling errors when starting a recording:

// Your implementation
private func handleRecording() async {
    do {
        let templates: [OutputFormatTemplate] = [
            // Your templates
        ]
        
        // Calling SDK function
        try await viewModel.startRecording(
            conversationType: VoiceConversationType.conversation,
            inputLanguage: [InputLanguageType.english, InputLanguageType.hindi],
            templates: templates,
            modelType: ModelType.lite
        )
        
        // Successfully started recording
    } catch let error as EkaScribeError {
        // Handle SDK-specific errors
        await MainActor.run {
            switch error {
            case .freeSessionLimitReached:
                // Show upgrade prompt
                showUpgradeAlert()
            case .microphonePermissionDenied:
                // Show permission alert
                showPermissionAlert()
            case .microphoneIsInUse:
                // Show microphone in use alert
                showMicrophoneInUseAlert()
            default:
                // Handle other SDK-specific errors
                showGenericError(error.localizedDescription)
            }
        }
    } catch {
        // Handle generic errors (network, API, etc.)
        await MainActor.run {
            showGenericError(error.localizedDescription)
        }
    }
}

API Reference

VoiceToRxViewModel

Properties

@Published public var screenState: RecordConsultationState

Methods

// Core recording methods
public func startRecording(
    conversationType: VoiceConversationType,
    inputLanguage: [InputLanguageType],
    templates: [OutputFormatTemplate],
    modelType: ModelType
) async throws

public func stopRecording() async throws
public func pauseRecording()
public func resumeRecording() throws

// Session management
public func deleteRecording(id: UUID)
public func clearSession()
public func retryIfNeeded()

VoiceToRxRepo

Template Management

public func getTemplates(completion: @escaping (Result<TemplateResponse, Error>) -> Void)
public func updateConfig(templates: [String], completion: @escaping (Result<Void, Error>) -> Void)

Session History

public func getEkaScribeHistory(completion: @escaping (Result<ScribeHistoryResponse, Error>) -> Void)

Result Management

public func fetchResultStatusResponse(
    sessionID: String,
    completion: @escaping (Result<VoiceToRxStatusResponse, Error>) -> Void
)
// Note: All response values are base64 encoded. Decode before displaying.

public func switchTemplate(
    templateID: String,
    sessionID: String,
    completion: @escaping (Result<VoiceToRxStatusResponse, Error>) -> Void
)

public func updateResult(
    sessionID: String,
    request: UpdateResultRequest,
    completion: @escaping (Result<Void, ErrorStatus>) -> Void
)

Configuration Classes

V2RxInitConfigurations

public class V2RxInitConfigurations {
    public static let shared: V2RxInitConfigurations
    
    public var modelContainer: ModelContainer?
    public var clientId: String?
    public var ownerName: String?
    public var ownerOID: String?
    public var ownerUUID: String?
    public var subOwnerOID: String?
    public var subOwnerName: String?
    public var appointmentID: String?
}

AuthTokenHolder

public class AuthTokenHolder {
    public static let shared: AuthTokenHolder
    
    public var authToken: String?
    public var refreshToken: String?
    public var bid: String?
}

Data Models

OutputFormatTemplate

public struct OutputFormatTemplate {
    public let templateID: String
    public let templateType: TemplateType
    public let templateName: String
    
    public init(
        templateID: String,
        templateType: TemplateType,
        templateName: String
    )
}

VoiceToRxStatusResponse

public struct VoiceToRxStatusResponse {
    public let data: VoiceToRxStatusData?
}

public struct VoiceToRxStatusData {
    public let templateResults: TemplateResults?
    public let output: [VoiceToRxOutput]?
    public let audioMatrix: AudioMatrix?
}

public struct TemplateResults {
    public let custom: [VoiceToRxOutput]?
    public let transcript: [VoiceToRxOutput]?
}

Best Practices

Memory Management

class RecordingViewController: UIViewController {
    private var viewModel: VoiceToRxViewModel?
    
    deinit {
        // Clean up resources
        viewModel?.clearSession()
    }
    
    override func viewWillDisappear(_ animated: Bool) {
        super.viewWillDisappear(animated)
        
        // Pause recording when leaving screen
        if case .listening = viewModel?.screenState {
            viewModel?.pauseRecording()
        }
    }
}

State Management

Always observe screenState changes to keep your UI in sync:

.onChange(of: viewModel.screenState) { oldState, newState in
    // Update UI based on state
    handleStateChange(newState)
}

Error Handling

Always handle errors from async operations using try-catch:

// Your implementation
private func startRecording() async {
    do {
        let templates: [OutputFormatTemplate] = [
            // Your templates
        ]
        
        // Calling SDK function
        try await viewModel.startRecording(
            conversationType: VoiceConversationType.conversation,
            inputLanguage: [InputLanguageType.english, InputLanguageType.hindi],
            templates: templates,
            modelType: ModelType.lite
        )
        
        // Successfully started recording
    } catch {
        // Handle error appropriately
        await MainActor.run {
            showError(error)
        }
    }
}

Troubleshooting

Microphone Permission Issues

If microphone permission is denied:

Check Info.plist has NSMicrophoneUsageDescription
Request permission programmatically if needed
Guide users to Settings if permission was previously denied

Session Creation Failures

If sessions fail to create:

Verify ownerOID is set in V2RxInitConfigurations.shared
Check AuthTokenHolder.shared.authToken is valid
Ensure network connectivity
Check for free session limit if applicable

Result Not Available

If results are not available after processing:

Check screenState transitions to .resultDisplay
Verify session ID is available: viewModel.sessionID
Use VoiceToRxRepo.shared.fetchResultStatusResponse to get full response
Check network connectivity for result retrieval

Name		Name	Last commit message	Last commit date
Latest commit History 361 Commits
Sources		Sources
.gitignore		.gitignore
Package.resolved		Package.resolved
Package.swift		Package.swift
README.md		README.md

eka-care/EkaVoiceToRx

Folders and files

Latest commit

History

Repository files navigation

EkaScribe iOS SDK

Overview

Getting Started

Key Features

Table of Contents

Requirements

System Permissions

Installation

Swift Package Manager

Package.swift

Import Statement

Integration Guide

Step 1: Configure Authentication and Doctor Information

Step 2: Initialize the View Model

Step 3: Start Recording

Parameters

Core Components

VoiceToRxViewModel

Recording States

Conversation Types

Model Types

Input Languages

Configuration

App-Level Setup

Session-Specific Configuration

Configuration Properties

Authentication

Recording Management

Starting Recording

Parameters

Pausing and Resuming

Stopping Recording

Handling Results

Template Management

Fetching Available Templates

Updating Favorite Templates

Using Templates in Recording

Session History

Fetching Session History

Result Management

Fetching Full Result Response

Switching Templates

Updating Results

Error Handling

Common Errors

Error Handling Example

API Reference

VoiceToRxViewModel

Properties

Methods

VoiceToRxRepo

Template Management

Session History

Result Management

Configuration Classes

V2RxInitConfigurations

AuthTokenHolder

Data Models

OutputFormatTemplate

VoiceToRxStatusResponse

Best Practices

Memory Management

State Management

Error Handling

Troubleshooting

Microphone Permission Issues

Session Creation Failures

Result Not Available

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases 23

Packages