rn-speech-to-text 0.1.0

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 tarun-vin
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -0,0 +1,2 @@
+# rn-speech-to-text
+speech-to-text package

package/android/build.gradle

@@ -0,0 +1,35 @@
+buildscript {
+    repositories {
+        mavenCentral()
+        google()
+    }
+}
+
+apply plugin: 'com.android.library'
+
+android {
+    namespace "com.vinfbsomni"
+    compileSdkVersion 34
+
+    defaultConfig {
+        minSdkVersion 21
+        targetSdkVersion 34
+    }
+
+    sourceSets {
+        main {
+            java.srcDirs = ['src/main/java']
+        }
+    }
+}
+
+repositories {
+    mavenCentral()
+    google()
+}
+
+dependencies {
+    implementation 'com.facebook.react:react-native:+'
+
+    implementation 'com.alphacephei:vosk-android:0.3.47'
+}

package/android/src/main/assets/models/vosk-model-small-en-us-0.15/README

@@ -0,0 +1,9 @@
+US English model for mobile Vosk applications
+
+Copyright 2020 Alpha Cephei Inc
+
+Accuracy: 10.38 (tedlium test) 9.85 (librispeech test-clean)
+Speed: 0.11xRT (desktop)
+Latency: 0.15s (right context)
+
+

package/android/src/main/assets/models/vosk-model-small-en-us-0.15/conf/mfcc.conf

@@ -0,0 +1,7 @@
+--sample-frequency=16000
+--use-energy=false
+--num-mel-bins=40
+--num-ceps=40
+--low-freq=20
+--high-freq=7600
+--allow-downsample=true

package/android/src/main/assets/models/vosk-model-small-en-us-0.15/conf/model.conf

@@ -0,0 +1,10 @@
+--min-active=200
+--max-active=3000
+--beam=10.0
+--lattice-beam=2.0
+--acoustic-scale=1.0
+--frame-subsampling-factor=3
+--endpoint.silence-phones=1:2:3:4:5:6:7:8:9:10
+--endpoint.rule2.min-trailing-silence=0.5
+--endpoint.rule3.min-trailing-silence=0.75
+--endpoint.rule4.min-trailing-silence=1.0

package/android/src/main/assets/models/vosk-model-small-en-us-0.15/graph/disambig_tid.int

@@ -0,0 +1,17 @@
+10015
+10016
+10017
+10018
+10019
+10020
+10021
+10022
+10023
+10024
+10025
+10026
+10027
+10028
+10029
+10030
+10031

package/android/src/main/assets/models/vosk-model-small-en-us-0.15/graph/phones/word_boundary.int

@@ -0,0 +1,166 @@
+1 nonword
+2 begin
+3 end
+4 internal
+5 singleton
+6 nonword
+7 begin
+8 end
+9 internal
+10 singleton
+11 begin
+12 end
+13 internal
+14 singleton
+15 begin
+16 end
+17 internal
+18 singleton
+19 begin
+20 end
+21 internal
+22 singleton
+23 begin
+24 end
+25 internal
+26 singleton
+27 begin
+28 end
+29 internal
+30 singleton
+31 begin
+32 end
+33 internal
+34 singleton
+35 begin
+36 end
+37 internal
+38 singleton
+39 begin
+40 end
+41 internal
+42 singleton
+43 begin
+44 end
+45 internal
+46 singleton
+47 begin
+48 end
+49 internal
+50 singleton
+51 begin
+52 end
+53 internal
+54 singleton
+55 begin
+56 end
+57 internal
+58 singleton
+59 begin
+60 end
+61 internal
+62 singleton
+63 begin
+64 end
+65 internal
+66 singleton
+67 begin
+68 end
+69 internal
+70 singleton
+71 begin
+72 end
+73 internal
+74 singleton
+75 begin
+76 end
+77 internal
+78 singleton
+79 begin
+80 end
+81 internal
+82 singleton
+83 begin
+84 end
+85 internal
+86 singleton
+87 begin
+88 end
+89 internal
+90 singleton
+91 begin
+92 end
+93 internal
+94 singleton
+95 begin
+96 end
+97 internal
+98 singleton
+99 begin
+100 end
+101 internal
+102 singleton
+103 begin
+104 end
+105 internal
+106 singleton
+107 begin
+108 end
+109 internal
+110 singleton
+111 begin
+112 end
+113 internal
+114 singleton
+115 begin
+116 end
+117 internal
+118 singleton
+119 begin
+120 end
+121 internal
+122 singleton
+123 begin
+124 end
+125 internal
+126 singleton
+127 begin
+128 end
+129 internal
+130 singleton
+131 begin
+132 end
+133 internal
+134 singleton
+135 begin
+136 end
+137 internal
+138 singleton
+139 begin
+140 end
+141 internal
+142 singleton
+143 begin
+144 end
+145 internal
+146 singleton
+147 begin
+148 end
+149 internal
+150 singleton
+151 begin
+152 end
+153 internal
+154 singleton
+155 begin
+156 end
+157 internal
+158 singleton
+159 begin
+160 end
+161 internal
+162 singleton
+163 begin
+164 end
+165 internal
+166 singleton

package/android/src/main/assets/models/vosk-model-small-en-us-0.15/ivector/global_cmvn.stats

@@ -0,0 +1,3 @@
+ [
+  1.682383e+11 -1.1595e+10 -1.521733e+10 4.32034e+09 -2.257938e+10 -1.969666e+10 -2.559265e+10 -1.535687e+10 -1.276854e+10 -4.494483e+09 -1.209085e+10 -5.64008e+09 -1.134847e+10 -3.419512e+09 -1.079542e+10 -4.145463e+09 -6.637486e+09 -1.11318e+09 -3.479773e+09 -1.245932e+08 -1.386961e+09 6.560655e+07 -2.436518e+08 -4.032432e+07 4.620046e+08 -7.714964e+07 9.551484e+08 -4.119761e+08 8.208582e+08 -7.117156e+08 7.457703e+08 -4.3106e+08 1.202726e+09 2.904036e+08 1.231931e+09 3.629848e+08 6.366939e+08 -4.586172e+08 -5.267629e+08 -3.507819e+08 1.679838e+09 
+  1.741141e+13 8.92488e+11 8.743834e+11 8.848896e+11 1.190313e+12 1.160279e+12 1.300066e+12 1.005678e+12 9.39335e+11 8.089614e+11 7.927041e+11 6.882427e+11 6.444235e+11 5.151451e+11 4.825723e+11 3.210106e+11 2.720254e+11 1.772539e+11 1.248102e+11 6.691599e+10 3.599804e+10 1.207574e+10 1.679301e+09 4.594778e+08 5.821614e+09 1.451758e+10 2.55803e+10 3.43277e+10 4.245286e+10 4.784859e+10 4.988591e+10 4.925451e+10 5.074584e+10 4.9557e+10 4.407876e+10 3.421443e+10 3.138606e+10 2.539716e+10 1.948134e+10 1.381167e+10 0 ]

package/android/src/main/assets/models/vosk-model-small-en-us-0.15/ivector/online_cmvn.conf

@@ -0,0 +1 @@
+# configuration file for apply-cmvn-online, used in the script ../local/run_online_decoding.sh

package/android/src/main/assets/models/vosk-model-small-en-us-0.15/ivector/splice.conf

@@ -0,0 +1,2 @@
+--left-context=3
+--right-context=3

package/android/src/main/assets/models/vosk-model-small-en-us-0.15/uuid

@@ -0,0 +1 @@
+custom-vosk-model-uuid-1

package/android/src/main/java/com/vinfbsomni/SpeechToTextModule.java

@@ -0,0 +1,310 @@
+package com.vinfbsomni;
+
+import android.Manifest;
+import android.content.pm.PackageManager;
+import android.media.AudioFormat;
+import android.media.AudioRecord;
+import android.media.MediaRecorder;
+import android.util.Log;
+
+import androidx.annotation.NonNull;
+import androidx.annotation.Nullable;
+import androidx.core.content.ContextCompat;
+
+import com.facebook.react.bridge.Arguments;
+import com.facebook.react.bridge.LifecycleEventListener;
+import com.facebook.react.bridge.Promise;
+import com.facebook.react.bridge.ReactApplicationContext;
+import com.facebook.react.bridge.ReactContextBaseJavaModule;
+import com.facebook.react.bridge.ReactMethod;
+import com.facebook.react.bridge.WritableMap;
+import com.facebook.react.modules.core.DeviceEventManagerModule;
+
+import org.json.JSONObject;
+import org.vosk.Model;
+import org.vosk.Recognizer;
+import org.vosk.android.StorageService;
+
+import java.io.IOException;
+
+public class SpeechToTextModule extends ReactContextBaseJavaModule implements LifecycleEventListener {
+
+  private static final String TAG = "SpeechToTextVosk";
+
+  // AudioRecord config (16k mono PCM16 is common for ASR)
+  private static final int SAMPLE_RATE = 16000;
+  private static final int CHANNEL_CONFIG = AudioFormat.CHANNEL_IN_MONO;
+  private static final int AUDIO_FORMAT = AudioFormat.ENCODING_PCM_16BIT;
+
+  // This is the folder under android/app/src/main/assets/
+  // Your model is at: android/app/src/main/assets/models/vosk-model-small-en-us-0.15/...
+  private static final String ASSET_MODEL_PATH = "models/vosk-model-small-en-us-0.15";
+  private static final String UNPACKED_MODEL_DIR = "model"; // directory name in app internal storage
+
+  private final ReactApplicationContext reactContext;
+
+  private AudioRecord audioRecord;
+  private boolean isRecording = false;
+  private Thread recordingThread;
+
+  private Model voskModel;
+  private Recognizer recognizer;
+
+  public SpeechToTextModule(ReactApplicationContext reactContext) {
+    super(reactContext);
+    this.reactContext = reactContext;
+    reactContext.addLifecycleEventListener(this);
+  }
+
+  @NonNull
+  @Override
+  public String getName() {
+    return "SpeechToText";
+  }
+
+  // ---- Helpers ----
+
+  private void sendEvent(String eventName, @Nullable WritableMap params) {
+    reactContext
+      .getJSModule(DeviceEventManagerModule.RCTDeviceEventEmitter.class)
+      .emit(eventName, params);
+  }
+
+  private void sendErrorEvent(String message) {
+    WritableMap map = Arguments.createMap();
+    map.putString("error", message);
+    sendEvent("onSpeechError", map);
+  }
+
+  // ---- Model initialization (must be called from JS before start()) ----
+
+  @ReactMethod
+  public void initModel(Promise promise) {
+    if (voskModel != null) {
+      // Already initialized
+      promise.resolve(null);
+      return;
+    }
+
+    try {
+      Log.d(TAG, "Unpacking Vosk model from assets: " + ASSET_MODEL_PATH);
+
+      StorageService.unpack(
+        reactContext,
+        ASSET_MODEL_PATH,
+        UNPACKED_MODEL_DIR,
+        (Model model) -> {
+          voskModel = model;
+          Log.d(TAG, "Vosk model loaded successfully");
+          promise.resolve(null);
+        },
+        (IOException e) -> {
+          Log.e(TAG, "Failed to unpack/load Vosk model", e);
+          promise.reject("model_init_failed", e);
+        }
+      );
+    } catch (Exception e) {
+      Log.e(TAG, "Exception during model init", e);
+      promise.reject("model_init_failed", e);
+    }
+  }
+
+  // ---- Authorization (status only; runtime request done in JS) ----
+
+  @ReactMethod
+  public void requestAuthorization(Promise promise) {
+    int status = ContextCompat.checkSelfPermission(
+      reactContext,
+      Manifest.permission.RECORD_AUDIO
+    );
+
+    if (status == PackageManager.PERMISSION_GRANTED) {
+      promise.resolve("authorized");
+    } else {
+      promise.resolve("denied");
+    }
+  }
+
+  // ---- Start raw audio capture with AudioRecord + Vosk ----
+
+  @ReactMethod
+  public void start(Promise promise) {
+    if (isRecording) {
+      // Already recording; just resolve
+      promise.resolve(null);
+      return;
+    }
+
+    int status = ContextCompat.checkSelfPermission(
+      reactContext,
+      Manifest.permission.RECORD_AUDIO
+    );
+    if (status != PackageManager.PERMISSION_GRANTED) {
+      promise.reject("mic_permission_denied", "RECORD_AUDIO permission not granted");
+      return;
+    }
+
+    if (voskModel == null) {
+      // JS must call initModel() first and wait for it to complete
+      promise.reject("model_not_initialized", "Call initModel() and wait for it to complete before calling start()");
+      return;
+    }
+
+    try {
+      // Create recognizer for this session
+      releaseRecognizer();
+      recognizer = new Recognizer(voskModel, SAMPLE_RATE);
+
+      int minBufferSize = AudioRecord.getMinBufferSize(
+        SAMPLE_RATE,
+        CHANNEL_CONFIG,
+        AUDIO_FORMAT
+      );
+      if (minBufferSize == AudioRecord.ERROR || minBufferSize == AudioRecord.ERROR_BAD_VALUE) {
+        promise.reject("audio_init_error", "Invalid buffer size for AudioRecord");
+        return;
+      }
+
+      audioRecord = new AudioRecord(
+        MediaRecorder.AudioSource.MIC,
+        SAMPLE_RATE,
+        CHANNEL_CONFIG,
+        AUDIO_FORMAT,
+        minBufferSize
+      );
+
+      if (audioRecord.getState() != AudioRecord.STATE_INITIALIZED) {
+        audioRecord.release();
+        audioRecord = null;
+        promise.reject("audio_init_error", "AudioRecord failed to initialize");
+        return;
+      }
+
+      isRecording = true;
+      audioRecord.startRecording();
+
+      // Background thread that continually reads audio & sends to Vosk
+      recordingThread = new Thread(() -> {
+        byte[] buffer = new byte[minBufferSize];
+
+        try {
+          while (isRecording && audioRecord != null && recognizer != null) {
+            int read = audioRecord.read(buffer, 0, buffer.length);
+            if (read > 0) {
+              boolean isFinal = recognizer.acceptWaveForm(buffer, read);
+
+              String resultJson;
+              if (isFinal) {
+                resultJson = recognizer.getResult();       // final segment
+              } else {
+                resultJson = recognizer.getPartialResult(); // partial segment
+              }
+
+              try {
+                JSONObject json = new JSONObject(resultJson);
+                if (isFinal) {
+                  String text = json.optString("text", "");
+                  if (!text.isEmpty()) {
+                    WritableMap map = Arguments.createMap();
+                    map.putString("value", text);
+                    sendEvent("onSpeechResults", map);
+                  }
+                } else {
+                  String partial = json.optString("partial", "");
+                  if (!partial.isEmpty()) {
+                    WritableMap map = Arguments.createMap();
+                    map.putString("value", partial);
+                    sendEvent("onSpeechPartialResults", map);
+                  }
+                }
+              } catch (Exception e) {
+                Log.e(TAG, "Vosk JSON parse error", e);
+              }
+            }
+          }
+        } catch (Exception e) {
+          Log.e(TAG, "AudioRecord / Vosk error", e);
+          sendErrorEvent("AudioRecord / Vosk error: " + e.getMessage());
+        }
+      }, "AudioRecordVoskThread");
+
+      recordingThread.start();
+      promise.resolve(null);
+    } catch (Exception e) {
+      Log.e(TAG, "start AudioRecord+Vosk error", e);
+      stopInternal();
+      promise.reject("start_error", e);
+    }
+  }
+
+  // ---- Stop raw audio capture ----
+
+  @ReactMethod
+  public void stop(Promise promise) {
+    stopInternal();
+    promise.resolve(null);
+  }
+
+  private void stopInternal() {
+    isRecording = false;
+
+    if (recordingThread != null) {
+      try {
+        recordingThread.join();
+      } catch (InterruptedException e) {
+        // ignore
+      }
+      recordingThread = null;
+    }
+
+    if (audioRecord != null) {
+      try {
+        if (audioRecord.getState() == AudioRecord.STATE_INITIALIZED) {
+          audioRecord.stop();
+        }
+      } catch (Exception e) {
+        Log.e(TAG, "AudioRecord stop error", e);
+      }
+      audioRecord.release();
+      audioRecord = null;
+    }
+
+    releaseRecognizer();
+    // Usually you keep the Model alive for the app lifetime to avoid reload cost.
+    // If you really want to free memory completely, uncomment below:
+    /*
+    if (voskModel != null) {
+      try {
+        voskModel.close();
+      } catch (Exception e) {
+        Log.e(TAG, "Error closing Vosk model", e);
+      }
+      voskModel = null;
+    }
+    */
+  }
+
+  private void releaseRecognizer() {
+    if (recognizer != null) {
+      try {
+        recognizer.close();
+      } catch (Exception e) {
+        Log.e(TAG, "Error closing recognizer", e);
+      }
+      recognizer = null;
+    }
+  }
+
+  // ---- Lifecycle cleanup ----
+
+  @Override
+  public void onHostResume() { }
+
+  @Override
+  public void onHostPause() { }
+
+  @Override
+  public void onHostDestroy() {
+    stopInternal();
+  }
+}

package/android/src/main/java/com/vinfbsomni/SpeechToTextPackage.java

@@ -0,0 +1,25 @@
+package com.vinfbsomni;
+
+import com.facebook.react.ReactPackage;
+import com.facebook.react.bridge.NativeModule;
+import com.facebook.react.bridge.ReactApplicationContext;
+import com.facebook.react.uimanager.ViewManager;
+
+import java.util.ArrayList;
+import java.util.Collections;
+import java.util.List;
+
+public class SpeechToTextPackage implements ReactPackage {
+
+  @Override
+  public List<NativeModule> createNativeModules(ReactApplicationContext reactContext) {
+    List<NativeModule> modules = new ArrayList<>();
+    modules.add(new SpeechToTextModule(reactContext));
+    return modules;
+  }
+
+  @Override
+  public List<ViewManager> createViewManagers(ReactApplicationContext reactContext) {
+    return Collections.emptyList();
+  }
+}

package/ios/SpeechToText.m

@@ -0,0 +1,22 @@
+//
+//  SpeechToText.m
+//  Sango
+//
+//  Created by TarunKumar on 26/02/26.
+//
+
+#import <React/RCTBridgeModule.h>
+#import <React/RCTEventEmitter.h>
+
+@interface RCT_EXTERN_MODULE(SpeechToText, RCTEventEmitter)
+
+RCT_EXTERN_METHOD(requestAuthorization:(RCTPromiseResolveBlock)resolve
+                  rejecter:(RCTPromiseRejectBlock)reject)
+
+RCT_EXTERN_METHOD(start:(RCTPromiseResolveBlock)resolve
+                  rejecter:(RCTPromiseRejectBlock)reject)
+
+RCT_EXTERN_METHOD(stop:(RCTPromiseResolveBlock)resolve
+                  rejecter:(RCTPromiseRejectBlock)reject)
+
+@end

package/ios/SpeechToText.swift

@@ -0,0 +1,138 @@
+//
+//  SpeechToText.swift
+//  Sango
+//
+//  Created by TarunKumar on 26/02/26.
+//
+
+import Foundation
+import Speech
+import AVFoundation
+import React
+
+@objc(SpeechToText)
+class SpeechToText: RCTEventEmitter {
+
+  private let audioEngine = AVAudioEngine()
+  private var speechRecognizer = SFSpeechRecognizer(locale: Locale(identifier: "en-US"))
+  private var recognitionRequest: SFSpeechAudioBufferRecognitionRequest?
+  private var recognitionTask: SFSpeechRecognitionTask?
+
+  // MARK: - React Native setup
+
+  override static func requiresMainQueueSetup() -> Bool {
+    return true
+  }
+
+  override func supportedEvents() -> [String]! {
+    return ["onSpeechResults", "onSpeechPartialResults", "onSpeechError"]
+  }
+
+  // MARK: - Public methods exposed to JS
+
+  @objc
+  func requestAuthorization(_ resolve: @escaping RCTPromiseResolveBlock,
+                            rejecter reject: @escaping RCTPromiseRejectBlock) {
+    SFSpeechRecognizer.requestAuthorization { status in
+      DispatchQueue.main.async {
+        switch status {
+        case .authorized:
+          resolve("authorized")
+        case .denied:
+          resolve("denied")
+        case .restricted:
+          resolve("restricted")
+        case .notDetermined:
+          resolve("notDetermined")
+        @unknown default:
+          resolve("unknown")
+        }
+      }
+    }
+  }
+
+  @objc
+  func start(_ resolve: @escaping RCTPromiseResolveBlock,
+             rejecter reject: @escaping RCTPromiseRejectBlock) {
+    do {
+      try startRecording()
+      resolve(nil)
+    } catch {
+      stopRecording()
+      reject("start_error", error.localizedDescription, error)
+      sendEvent(withName: "onSpeechError", body: ["error": error.localizedDescription])
+    }
+  }
+
+  @objc
+  func stop(_ resolve: @escaping RCTPromiseResolveBlock,
+            rejecter reject: @escaping RCTPromiseRejectBlock) {
+    stopRecording()
+    resolve(nil)
+  }
+
+  // MARK: - Recording logic
+
+  private func startRecording() throws {
+    // reset any previous session
+    stopRecording()
+
+    // Configure audio session
+    let audioSession = AVAudioSession.sharedInstance()
+    try audioSession.setCategory(.record, mode: .measurement, options: .duckOthers)
+    try audioSession.setActive(true, options: .notifyOthersOnDeactivation)
+
+    recognitionRequest = SFSpeechAudioBufferRecognitionRequest()
+    guard let recognitionRequest = recognitionRequest else {
+      throw NSError(
+        domain: "SpeechToText",
+        code: -1,
+        userInfo: [NSLocalizedDescriptionKey: "Unable to create recognition request"]
+      )
+    }
+    recognitionRequest.shouldReportPartialResults = true
+
+    let inputNode = audioEngine.inputNode
+
+    recognitionTask = speechRecognizer?.recognitionTask(with: recognitionRequest) {
+      [weak self] result, error in
+      guard let self = self else { return }
+
+      if let result = result {
+        let transcript = result.bestTranscription.formattedString
+        let eventName = result.isFinal ? "onSpeechResults" : "onSpeechPartialResults"
+        self.sendEvent(withName: eventName, body: ["value": transcript])
+        // NOTE: do NOT auto-stop here; JS controls stop()
+      }
+
+      if let error = error {
+        self.sendEvent(withName: "onSpeechError", body: ["error": error.localizedDescription])
+        self.stopRecording()
+      }
+    }
+
+    let recordingFormat = inputNode.outputFormat(forBus: 0)
+    inputNode.removeTap(onBus: 0)
+    inputNode.installTap(onBus: 0, bufferSize: 1024, format: recordingFormat) {
+      [weak self] buffer, _ in
+      self?.recognitionRequest?.append(buffer)
+    }
+
+    audioEngine.prepare()
+    try audioEngine.start()
+  }
+
+  private func stopRecording() {
+    if audioEngine.isRunning {
+      audioEngine.stop()
+      audioEngine.inputNode.removeTap(onBus: 0)
+    }
+
+    recognitionRequest?.endAudio()
+    recognitionTask?.cancel()
+    recognitionRequest = nil
+    recognitionTask = nil
+
+    try? AVAudioSession.sharedInstance().setActive(false)
+  }
+}

package/package.json

@@ -0,0 +1,37 @@
+{
+  "name": "rn-speech-to-text",
+  "version": "0.1.0",
+  "description": "Speech-to-text native module for React Native (iOS & Android).",
+  "main": "src/index.js",
+  "react-native": "src/index.js",
+  "keywords": [
+    "react-native",
+    "speech-to-text",
+    "speech",
+    "voice",
+    "stt"
+  ],
+  "author": "Tarun Kumar <tarun.kumar@vinculumgroup.com> (https://github.com/tarun-vin)",
+  "license": "MIT",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/tarun-vin/rn-speech-to-text.git"
+  },
+  "bugs": {
+    "url": "https://github.com/tarun-vin/rn-speech-to-text/issues"
+  },
+  "homepage": "https://github.com/tarun-vin/rn-speech-to-text#readme",
+  "peerDependencies": {
+    "react": ">=18.0.0",
+    "react-native": ">=0.71.0"
+  },
+  "files": [
+    "src",
+    "android",
+    "ios",
+    "react-native.config.js",
+    "rn-speech-to-text.podspec",
+    "README.md",
+    "LICENSE"
+  ]
+}

package/react-native.config.js

@@ -0,0 +1 @@
+module.exports = {};

package/rn-speech-to-text.podspec

@@ -0,0 +1,23 @@
+require 'json'
+
+package = JSON.parse(File.read(File.join(__dir__, 'package.json')))
+
+Pod::Spec.new do |s|
+  s.name         = "rn-speech-to-text"
+  s.version      = package['version']
+  s.summary      = "Speech-to-text native module for React Native (iOS & Android)."
+  s.description  = <<-DESC
+                   Speech-to-text native module for React Native using iOS Speech framework and Android SpeechRecognizer.
+                   DESC
+  s.homepage     = package['homepage']
+  s.license      = { :type => package['license'] || "MIT" }
+  s.author       = package['author'] || { "Tarun Kumar" => "tarun.kumar@vinculumgroup.com" }
+
+  s.platform     = :ios, "11.0"
+  s.source = { :git => "https://github.com/tarun-vin/rn-speech-to-text.git", :tag => "0.1.0" }
+
+  s.source_files = "ios/**/*.{h,m,mm,swift}"
+  s.requires_arc = true
+
+  s.dependency "React-Core"
+end

package/src/SpeechToText.js

@@ -0,0 +1,134 @@
+// SpeechToText.js
+import {
+  NativeModules,
+  NativeEventEmitter,
+  Platform,
+  PermissionsAndroid,
+} from 'react-native';
+
+const { SpeechToText } = NativeModules;
+
+if (!SpeechToText) {
+  throw new Error('SpeechToText native module not linked');
+}
+
+const emitter = new NativeEventEmitter(SpeechToText);
+
+// Android-only transcript buffers (used to merge final + partial)
+let androidFinalTranscript = '';
+let androidLivePartial = '';
+
+function normalizeText(value) {
+  return typeof value === 'string' ? value.trim() : '';
+}
+
+function buildAndroidMergedText() {
+  return [androidFinalTranscript, androidLivePartial].filter(Boolean).join(' ').trim();
+}
+
+function resetAndroidBuffers() {
+  androidFinalTranscript = '';
+  androidLivePartial = '';
+}
+
+export async function requestAuthorization() {
+  if (Platform.OS === 'ios') {
+    // iOS: use native Speech framework auth
+    return SpeechToText.requestAuthorization();
+  }
+
+  // Android: request RECORD_AUDIO runtime permission
+  const result = await PermissionsAndroid.request(
+    PermissionsAndroid.PERMISSIONS.RECORD_AUDIO,
+    {
+      title: 'Microphone Permission',
+      message: 'We need access to your microphone for voice search.',
+      buttonPositive: 'OK',
+    },
+  );
+
+  if (result === PermissionsAndroid.RESULTS.GRANTED) {
+    return 'authorized';
+  }
+  return 'denied';
+}
+
+export function initModel() {
+  if (Platform.OS === 'ios') return;
+  return SpeechToText.initModel();
+}
+
+export function start() {
+  if (Platform.OS === 'android') {
+    resetAndroidBuffers();
+  }
+  return SpeechToText.start();
+}
+
+export async function stop() {
+  try {
+    return await SpeechToText.stop();
+  } finally {
+    // Prepare for next session
+    if (Platform.OS === 'android') {
+      resetAndroidBuffers();
+    }
+  }
+}
+
+export function addResultListener(listener) {
+  // Keep raw final events behavior unchanged
+  return emitter.addListener('onSpeechResults', listener);
+}
+
+export function addPartialResultListener(listener) {
+  if (Platform.OS !== 'android') {
+    return emitter.addListener('onSpeechPartialResults', listener);
+  }
+
+  // Android:
+  // - partial events update current segment
+  // - final events commit segment
+  // - both emit merged text via onSpeechPartialResults listener so existing UI code works unchanged
+  const emitMerged = (event = {}) => {
+    listener({
+      ...event,
+      value: buildAndroidMergedText(),
+    });
+  };
+
+  const partialSub = emitter.addListener('onSpeechPartialResults', (event) => {
+    androidLivePartial = normalizeText(event?.value);
+    emitMerged(event);
+  });
+
+  const finalSub = emitter.addListener('onSpeechResults', (event) => {
+    const finalSegment = normalizeText(event?.value);
+
+    if (finalSegment) {
+      if (!androidFinalTranscript) {
+        androidFinalTranscript = finalSegment;
+      } else if (
+        androidFinalTranscript !== finalSegment &&
+        !androidFinalTranscript.endsWith(` ${finalSegment}`)
+      ) {
+        androidFinalTranscript = `${androidFinalTranscript} ${finalSegment}`;
+      }
+    }
+
+    // Clear current in-progress partial once finalized
+    androidLivePartial = '';
+    emitMerged(event);
+  });
+
+  return {
+    remove() {
+      partialSub.remove();
+      finalSub.remove();
+    },
+  };
+}
+
+export function addErrorListener(listener) {
+  return emitter.addListener('onSpeechError', listener);
+}

package/src/index.js

@@ -0,0 +1 @@
+export * from './SpeechToText';