termux-stt 1.4.0 → 2.0.0

package/README.md

@@ -1,117 +1,84 @@
-# Termux-STT
-
-[![PyPI](https://img.shields.io/pypi/v/termux-stt.svg?style=flat-square&color=0369a1)](https://pypi.org/project/termux-stt/)
-[![Python](https://img.shields.io/pypi/pyversions/termux-stt.svg?style=flat-square)](https://pypi.org/project/termux-stt/)
-[![npm](https://img.shields.io/npm/v/termux-stt.svg?style=flat-square&color=b91c1c)](https://www.npmjs.com/package/termux-stt)
-[![npm downloads](https://img.shields.io/npm/dm/termux-stt.svg?style=flat-square&color=b91c1c)](https://www.npmjs.com/package/termux-stt)
-[![License](https://img.shields.io/badge/License-Apache_2.0-004499.svg?style=flat-square)](https://github.com/uno-km/termux-stt)
-
-> **디바이스 리소스를 활용한 통합 온디바이스 음성인식(STT) 및 순수 Python 128d X-Vector 화자 분리 프레임워크**  
-> *Unified On-Device Speech-to-Text Utilizing Device Resources & Pure Python 128d X-Vector Speaker Diarization*
-
----
-
-## Architecture & Overview
-
-Whisper.cpp, Vosk, Sherpa-ONNX 3대 네이티브 엔진을 통합하고, 닫힌 형태 순수 Python 128차원 X-Vector 클러스터링을 결합하여 80MB 미만의 초경량 메모리로 100% 온디바이스 실시간 음성인식과 화자 분리를 구현합니다.
-
-Integrates Whisper.cpp, Vosk, and Sherpa-ONNX with a closed-form pure-Python 128-dimensional X-Vector clustering algorithm that operates in under 80MB RAM with zero cloud egress.
-
----
-
-## Installation & Quickstart
-
-### Python (PyPI)
-```bash
-pip install termux-stt
-```
-```python
-from termux_stt import create_engine
-
-# 1. Initialize Engine with Hybrid GPU-Encoder / CPU-Decoder Acceleration
-engine = create_engine("whisper", model="small", lang="en", threads=4, split_mode=True)
-
-# 2. Transcribe Audio directly into Subtitles
-result = engine.transcribe("samples/jfk_1min.wav")
-print("Transcript:\n", result.text)
-print("SRT Subtitles:\n", result.to_srt())
-
-# 3. 2-Speaker Diarization without PyTorch
-hybrid = create_engine("hybrid", lang="en", num_speakers=2)
-diar_result = hybrid.diarize("samples/jfk_1min.wav")
-for seg in diar_result.segments:
-    print(f"[{seg.speaker}] ({seg.start:.1f}s -> {seg.end:.1f}s): {seg.text}")
-
-```
-
-### Node.js / TypeScript (npm)
-```bash
-npm install termux-stt
-```
-```typescript
-const { createEngine } = require("termux-stt");
-
-async function main() {
-  // 1. Initialize Whisper Engine
-  const engine = createEngine("whisper", { model: "tiny", lang: "en", threads: 4 });
-
-  // 2. Transcribe Audio
-  const result = await engine.transcribe("samples/jfk_1min.wav");
-  console.log("Transcript:", result.text);
-  console.log("SRT Subtitles:\n", result.toSrt());
-}
-main();
-
-```
-
----
-
-## Distributed Clustering & Memory Pooling (AMEVA-Cluster)
-
-Termux-STT natively integrates with **AMEVA-Cluster** (`pip install ameva-cluster`) for distributed large-model speech recognition (e.g. `whisper-large-v3-turbo`) across interconnected mobile fleets.
-
-### 1. Install Cluster Runtime
-```bash
-pip install ameva-cluster
-# or Node.js:
-npm install @ameva/cluster
-```
-
-### 2. Launch Worker Node on Remote Phone
-```bash
-# On remote worker device (e.g. Galaxy A53):
-ameva-cluster worker --port 50052
-```
-
-### 3. Distributed Transcription via Master Node
-```bash
-# Master node sharding Whisper-Large layers across remote phone RAM:
-termux-stt transcribe meeting.wav \
-  --model whisper-large-v3-turbo.bin \
-  --rpc 192.0.2.10:50052,192.0.2.11:50052
-```
-
-```python
-from termux_stt import create_engine
-
-# Python SDK Distributed STT
-engine = create_engine(
-    "whisper",
-    model="large-v3-turbo",
-    cluster_rpc_servers="192.0.2.10:50052,192.0.2.11:50052"
-)
-result = engine.transcribe("meeting.wav")
-print(result.text)
-```
-
----
-
-## Official Documentation & Benchmarks
-- [Official Architecture & API Reference](https://uno-km.vercel.app/lib/stt/)
-- [Ecosystem Metrics & Registry Stats](https://uno-km.vercel.app/foundation/metrics)
-- [AMEVA Open-Source Foundation Portal](https://uno-km.vercel.app/foundation/index.html)
-
----
-
-## License
-Licensed under the Apache-2.0 License. Copyright (c) 2026 Eunho Kim ([@uno-km](https://github.com/uno-km)).
+# Termux-STT
+
+[![PyPI](https://img.shields.io/pypi/v/termux-stt.svg?style=flat-square&color=0369a1)](https://pypi.org/project/termux-stt/)
+[![Python](https://img.shields.io/pypi/pyversions/termux-stt.svg?style=flat-square)](https://pypi.org/project/termux-stt/)
+[![npm](https://img.shields.io/npm/v/termux-stt.svg?style=flat-square&color=b91c1c)](https://www.npmjs.com/package/termux-stt)
+[![npm downloads](https://img.shields.io/npm/dm/termux-stt.svg?style=flat-square&color=b91c1c)](https://www.npmjs.com/package/termux-stt)
+[![License](https://img.shields.io/badge/License-Apache_2.0-004499.svg?style=flat-square)](https://github.com/uno-km/termux-stt)
+
+> **디바이스 리소스를 활용한 통합 온디바이스 음성인식(STT) 및 차세대 신경망 화자 분리(PyAnnote 3.0 + CAM++ 192d + TS-VAD) 프레임워크**  
+> *Unified On-Device Speech-to-Text & Neural Multi-Speaker Diarization (PyAnnote 3.0 + CAM++ 192d + TS-VAD Overlap Resolver)*
+
+---
+
+## Architecture & Overview
+
+Whisper.cpp(네이티브 Vulkan GPU / NEON CPU) 및 Sherpa-ONNX 듀얼 모던 엔진을 통합하고, PyAnnote 3.0 신경망 세그멘테이션, 3D-Speaker CAM++ 192차원 성문 임베딩, TS-VAD 동시 발화(마이크 물림) 분리기를 결합하여 완전 온디바이스 실시간 다중 화자 음성인식을 구현합니다.
+
+Integrates Whisper.cpp (Native Vulkan GPU / NEON CPU) and Sherpa-ONNX with PyAnnote 3.0 neural segmentation, 3D-Speaker CAM++ 192-dim embeddings, and TS-VAD overlapped speech resolution with zero cloud egress.
+
+---
+
+## Installation & Quickstart
+
+### Python (PyPI)
+```bash
+pip install termux-stt
+```
+```python
+from termux_stt import create_engine
+
+# 1. Initialize Whisper Engine with Vulkan GPU / CPU Hybrid Acceleration
+engine = create_engine("whisper", model="small", lang="ko", threads=4, split_mode=True)
+
+# 2. Transcribe Audio directly into Subtitles
+result = engine.transcribe("samples/jfk_1min.wav")
+print("Transcript:\n", result.text)
+print("SRT Subtitles:\n", result.to_srt())
+
+# 3. Multi-Speaker Neural Diarization (PyAnnote 3.0 + CAM++ 192d + TS-VAD)
+hybrid = create_engine("hybrid", lang="ko", num_speakers=2)
+diar_result = hybrid.diarize("samples/kor_영어로화자분리.wav")
+for seg in diar_result.segments:
+    print(f"[{seg.speaker}] ({seg.start:.1f}s -> {seg.end:.1f}s): {seg.text}")
+
+```
+
+### Node.js / TypeScript (npm)
+```bash
+npm install termux-stt
+```
+```typescript
+const { createEngine } = require("termux-stt");
+
+async function main() {
+  // 1. Initialize Whisper Engine
+  const engine = createEngine("whisper", { model: "tiny", lang: "ko", threads: 4 });
+
+  // 2. Transcribe Audio
+  const result = await engine.transcribe("samples/jfk_1min.wav");
+  console.log("Transcript:", result.text);
+  console.log("SRT Subtitles:\n", result.toSrt());
+
+  // 3. Neural Diarization via Hybrid Engine
+  const hybrid = createEngine("hybrid", { lang: "ko", numSpeakers: 2 });
+  const diarResult = await hybrid.diarize("samples/kor_영어로화자분리.wav");
+  for (const seg of diarResult.segments) {
+    console.log(`[${seg.speaker}] (${seg.start.toFixed(1)}s -> ${seg.end.toFixed(1)}s): ${seg.text}`);
+  }
+}
+main();
+
+```
+
+---
+
+## Official Documentation & Benchmarks
+- [Official Architecture & API Reference](https://uno-km.vercel.app/lib/stt/)
+- [Ecosystem Metrics & Registry Stats](https://uno-km.vercel.app/foundation/metrics)
+- [AMEVA Open-Source Foundation Portal](https://uno-km.vercel.app/foundation/index.html)
+
+---
+
+## License
+Licensed under the Apache-2.0 License. Copyright (c) 2026 Eunho Kim ([@uno-km](https://github.com/uno-km)).

package/README.pypi.md

@@ -1,44 +1,44 @@
-# Termux-STT (Python)
-
-[![PyPI](https://img.shields.io/pypi/v/termux-stt.svg?style=flat-square&color=0369a1)](https://pypi.org/project/termux-stt/)
-[![Python](https://img.shields.io/pypi/pyversions/termux-stt.svg?style=flat-square)](https://pypi.org/project/termux-stt/)
-[![License](https://img.shields.io/badge/License-Apache_2.0-004499.svg?style=flat-square)](https://github.com/uno-km/termux-stt)
-
-> Unified On-Device Speech-to-Text Utilizing Device Resources & Pure Python 128d X-Vector Speaker Diarization
-
-## Installation
-
-```bash
-pip install termux-stt
-```
-
-## Quickstart
-
-```python
-from termux_stt import create_engine
-
-# 1. Initialize Engine with Hybrid GPU-Encoder / CPU-Decoder Acceleration
-engine = create_engine("whisper", model="small", lang="en", threads=4, split_mode=True)
-
-# 2. Transcribe Audio directly into Subtitles
-result = engine.transcribe("samples/jfk_1min.wav")
-print("Transcript:\n", result.text)
-print("SRT Subtitles:\n", result.to_srt())
-
-# 3. 2-Speaker Diarization without PyTorch
-hybrid = create_engine("hybrid", lang="en", num_speakers=2)
-diar_result = hybrid.diarize("samples/jfk_1min.wav")
-for seg in diar_result.segments:
-    print(f"[{seg.speaker}] ({seg.start:.1f}s -> {seg.end:.1f}s): {seg.text}")
-
-```
-
-## Description
-Integrates Whisper.cpp, Vosk, and Sherpa-ONNX with a closed-form pure-Python 128-dimensional X-Vector clustering algorithm that operates in under 80MB RAM with zero cloud egress.
-
-## Documentation
-- [Official Documentation & API Reference](https://uno-km.vercel.app/lib/stt/)
-- [GitHub Repository](https://github.com/uno-km/termux-stt)
-
-## License
-Apache-2.0 License. Copyright (c) 2026 Eunho Kim (@uno-km).
+# Termux-STT (Python)
+
+[![PyPI](https://img.shields.io/pypi/v/termux-stt.svg?style=flat-square&color=0369a1)](https://pypi.org/project/termux-stt/)
+[![Python](https://img.shields.io/pypi/pyversions/termux-stt.svg?style=flat-square)](https://pypi.org/project/termux-stt/)
+[![License](https://img.shields.io/badge/License-Apache_2.0-004499.svg?style=flat-square)](https://github.com/uno-km/termux-stt)
+
+> Unified On-Device Speech-to-Text & Neural Multi-Speaker Diarization (PyAnnote 3.0 + CAM++ 192d + TS-VAD Overlap Resolver)
+
+## Installation
+
+```bash
+pip install termux-stt
+```
+
+## Quickstart
+
+```python
+from termux_stt import create_engine
+
+# 1. Initialize Whisper Engine with Vulkan GPU / CPU Hybrid Acceleration
+engine = create_engine("whisper", model="small", lang="ko", threads=4, split_mode=True)
+
+# 2. Transcribe Audio directly into Subtitles
+result = engine.transcribe("samples/jfk_1min.wav")
+print("Transcript:\n", result.text)
+print("SRT Subtitles:\n", result.to_srt())
+
+# 3. Multi-Speaker Neural Diarization (PyAnnote 3.0 + CAM++ 192d + TS-VAD)
+hybrid = create_engine("hybrid", lang="ko", num_speakers=2)
+diar_result = hybrid.diarize("samples/kor_영어로화자분리.wav")
+for seg in diar_result.segments:
+    print(f"[{seg.speaker}] ({seg.start:.1f}s -> {seg.end:.1f}s): {seg.text}")
+
+```
+
+## Description
+Integrates Whisper.cpp (Native Vulkan GPU / NEON CPU) and Sherpa-ONNX with PyAnnote 3.0 neural segmentation, 3D-Speaker CAM++ 192-dim embeddings, and TS-VAD overlapped speech resolution with zero cloud egress.
+
+## Documentation
+- [Official Documentation & API Reference](https://uno-km.vercel.app/lib/stt/)
+- [GitHub Repository](https://github.com/uno-km/termux-stt)
+
+## License
+Apache-2.0 License. Copyright (c) 2026 Eunho Kim (@uno-km).

package/bin/cli.js

@@ -1,51 +1,51 @@
-#!/usr/bin/env node
-/**
- * AMEVA Standard Node.js CLI Runner for termux_stt.
- * Automatically resolves Python 3 environment and dispatches to python -m termux_stt.
- */
-const fs = require('fs');
-const { spawn, execSync } = require('child_process');
-
-function findPython() {
-  if (process.env.PYTHON && fs.existsSync(process.env.PYTHON)) {
-    return process.env.PYTHON;
-  }
-  const termuxBin = '/data/data/com.termux/files/usr/bin/python3';
-  if (fs.existsSync(termuxBin)) {
-    return termuxBin;
-  }
-  const termuxBinAlt = '/data/data/com.termux/files/usr/bin/python';
-  if (fs.existsSync(termuxBinAlt)) {
-    return termuxBinAlt;
-  }
-  const candidates = ['python3', 'python'];
-  for (const cmd of candidates) {
-    try {
-      const checkCmd = process.platform === 'win32' ? `where ${cmd}` : `command -v ${cmd}`;
-      const res = execSync(checkCmd, { stdio: ['ignore', 'pipe', 'ignore'] }).toString().trim();
-      if (res) return cmd;
-    } catch (_) {}
-  }
-  return 'python3';
-}
-
-const pythonBin = findPython();
-const args = ['-m', 'termux_stt', ...process.argv.slice(2)];
-
-const child = spawn(pythonBin, args, {
-  stdio: 'inherit',
-  env: process.env
-});
-
-child.on('error', (err) => {
-  console.error(`[${'termux_stt'}] Failed to spawn python process (${pythonBin}):`, err.message);
-  process.exit(1);
-});
-
-child.on('exit', (code, signal) => {
-  if (signal) {
-    process.kill(process.pid, signal);
-  } else {
-    process.exit(code || 0);
-  }
-});
+#!/usr/bin/env node
+/**
+ * AMEVA Standard Node.js CLI Runner for termux_stt.
+ * Automatically resolves Python 3 environment and dispatches to python -m termux_stt.
+ */
+const fs = require('fs');
+const { spawn, execSync } = require('child_process');
+
+function findPython() {
+  if (process.env.PYTHON && fs.existsSync(process.env.PYTHON)) {
+    return process.env.PYTHON;
+  }
+  const termuxBin = '/data/data/com.termux/files/usr/bin/python3';
+  if (fs.existsSync(termuxBin)) {
+    return termuxBin;
+  }
+  const termuxBinAlt = '/data/data/com.termux/files/usr/bin/python';
+  if (fs.existsSync(termuxBinAlt)) {
+    return termuxBinAlt;
+  }
+  const candidates = ['python3', 'python'];
+  for (const cmd of candidates) {
+    try {
+      const checkCmd = process.platform === 'win32' ? `where ${cmd}` : `command -v ${cmd}`;
+      const res = execSync(checkCmd, { stdio: ['ignore', 'pipe', 'ignore'] }).toString().trim();
+      if (res) return cmd;
+    } catch (_) {}
+  }
+  return 'python3';
+}
+
+const pythonBin = findPython();
+const args = ['-m', 'termux_stt', ...process.argv.slice(2)];
+
+const child = spawn(pythonBin, args, {
+  stdio: 'inherit',
+  env: process.env
+});
+
+child.on('error', (err) => {
+  console.error(`[${'termux_stt'}] Failed to spawn python process (${pythonBin}):`, err.message);
+  process.exit(1);
+});
+
+child.on('exit', (code, signal) => {
+  if (signal) {
+    process.kill(process.pid, signal);
+  } else {
+    process.exit(code || 0);
+  }
+});

package/bin/termux-stt.js

@@ -1,51 +1,51 @@
-#!/usr/bin/env node
-/**
- * AMEVA Standard Node.js CLI Runner for termux_stt.
- * Automatically resolves Python 3 environment and dispatches to python -m termux_stt.
- */
-const fs = require('fs');
-const { spawn, execSync } = require('child_process');
-
-function findPython() {
-  if (process.env.PYTHON && fs.existsSync(process.env.PYTHON)) {
-    return process.env.PYTHON;
-  }
-  const termuxBin = '/data/data/com.termux/files/usr/bin/python3';
-  if (fs.existsSync(termuxBin)) {
-    return termuxBin;
-  }
-  const termuxBinAlt = '/data/data/com.termux/files/usr/bin/python';
-  if (fs.existsSync(termuxBinAlt)) {
-    return termuxBinAlt;
-  }
-  const candidates = ['python3', 'python'];
-  for (const cmd of candidates) {
-    try {
-      const checkCmd = process.platform === 'win32' ? `where ${cmd}` : `command -v ${cmd}`;
-      const res = execSync(checkCmd, { stdio: ['ignore', 'pipe', 'ignore'] }).toString().trim();
-      if (res) return cmd;
-    } catch (_) {}
-  }
-  return 'python3';
-}
-
-const pythonBin = findPython();
-const args = ['-m', 'termux_stt.platform.installer', ...process.argv.slice(2)];
-
-const child = spawn(pythonBin, args, {
-  stdio: 'inherit',
-  env: process.env
-});
-
-child.on('error', (err) => {
-  console.error(`[${'termux_stt'}] Failed to spawn python process (${pythonBin}):`, err.message);
-  process.exit(1);
-});
-
-child.on('exit', (code, signal) => {
-  if (signal) {
-    process.kill(process.pid, signal);
-  } else {
-    process.exit(code || 0);
-  }
-});
+#!/usr/bin/env node
+/**
+ * AMEVA Standard Node.js CLI Runner for termux_stt.
+ * Automatically resolves Python 3 environment and dispatches to python -m termux_stt.
+ */
+const fs = require('fs');
+const { spawn, execSync } = require('child_process');
+
+function findPython() {
+  if (process.env.PYTHON && fs.existsSync(process.env.PYTHON)) {
+    return process.env.PYTHON;
+  }
+  const termuxBin = '/data/data/com.termux/files/usr/bin/python3';
+  if (fs.existsSync(termuxBin)) {
+    return termuxBin;
+  }
+  const termuxBinAlt = '/data/data/com.termux/files/usr/bin/python';
+  if (fs.existsSync(termuxBinAlt)) {
+    return termuxBinAlt;
+  }
+  const candidates = ['python3', 'python'];
+  for (const cmd of candidates) {
+    try {
+      const checkCmd = process.platform === 'win32' ? `where ${cmd}` : `command -v ${cmd}`;
+      const res = execSync(checkCmd, { stdio: ['ignore', 'pipe', 'ignore'] }).toString().trim();
+      if (res) return cmd;
+    } catch (_) {}
+  }
+  return 'python3';
+}
+
+const pythonBin = findPython();
+const args = ['-m', 'termux_stt.platform.installer', ...process.argv.slice(2)];
+
+const child = spawn(pythonBin, args, {
+  stdio: 'inherit',
+  env: process.env
+});
+
+child.on('error', (err) => {
+  console.error(`[${'termux_stt'}] Failed to spawn python process (${pythonBin}):`, err.message);
+  process.exit(1);
+});
+
+child.on('exit', (code, signal) => {
+  if (signal) {
+    process.kill(process.pid, signal);
+  } else {
+    process.exit(code || 0);
+  }
+});

package/index.d.ts

@@ -45,7 +45,7 @@ export class WhisperEngine extends Engine {
   diarize(audioPath: string, numSpeakers?: number): Promise<TranscriptResult>;
 }
 
-export class VoskEngine extends Engine {
+export class SherpaEngine extends Engine {
   transcribe(audioPath: string, options?: any): Promise<TranscriptResult>;
   diarize(audioPath: string, numSpeakers?: number): Promise<TranscriptResult>;
 }
@@ -56,8 +56,13 @@ export class HybridEngine extends Engine {
 }
 
 export function createEngine(
-  engineName?: 'whisper' | 'vosk' | 'hybrid' | string,
+  engineName?: 'whisper' | 'sherpa' | 'hybrid' | string,
   options?: EngineOptions
 ): Engine;
 
-export function formatTime(seconds: number, separator?: string): string;
+export class TermuxSTT {
+  engine: Engine;
+  constructor(options?: EngineOptions & { engine?: string });
+  transcribe(filePath: string, options?: any): Promise<TranscriptResult>;
+  diarize(filePath: string, options?: any): Promise<TranscriptResult>;
+}

package/index.js

@@ -1,10 +1,10 @@
 /**
- * termux-stt Node.js entry point
+ * termux-stt Node.js entry point (v2.0.0)
  */
 
 const { Engine, TranscriptResult, Segment, formatTime } = require('./lib/engine');
 const { WhisperEngine } = require('./lib/whisper');
-const { VoskEngine } = require('./lib/vosk');
+const { SherpaEngine } = require('./lib/sherpa');
 const { HybridEngine } = require('./lib/hybrid');
 
 function createEngine(engineName = 'whisper', options = {}) {
@@ -12,12 +12,14 @@ function createEngine(engineName = 'whisper', options = {}) {
   switch (name) {
     case 'whisper':
       return new WhisperEngine(options);
-    case 'vosk':
-      return new VoskEngine(options);
+    case 'sherpa':
+      return new SherpaEngine(options);
     case 'hybrid':
       return new HybridEngine(options);
+    case 'vosk':
+      throw new Error("Vosk has been deprecated and completely removed in v2.0.0. Please use 'sherpa' (SenseVoice) or 'whisper'.");
     default:
-      throw new Error(`Unknown engine: ${engineName}. Available: whisper, vosk, hybrid`);
+      throw new Error(`Unknown engine: ${engineName}. Available: whisper, sherpa, hybrid`);
   }
 }
 
@@ -41,7 +43,7 @@ module.exports = {
   TranscriptResult,
   Segment,
   WhisperEngine,
-  VoskEngine,
+  SherpaEngine,
   HybridEngine,
   formatTime
 };

package/lib/hybrid.js

@@ -53,7 +53,7 @@ class HybridEngine extends Engine {
 
   getInfo() {
     return {
-      name: 'Hybrid Vosk+Whisper (Node.js)',
+      name: 'Hybrid Neural Diarizer + Whisper (Node.js)',
       model: this.model,
       language: this.lang,
       numSpeakers: this.numSpeakers

package/lib/sherpa.js

@@ -0,0 +1,97 @@
+/**
+ * SherpaEngine for Node.js - Non-Autoregressive SenseVoice STT & Neural Diarization
+ */
+
+const { spawn } = require('child_process');
+const { Engine, TranscriptResult, Segment } = require('./engine');
+
+class SherpaEngine extends Engine {
+  constructor(config = {}) {
+    super(config);
+    this.model = config.model || 'sensevoice-small-int8';
+    this.lang = config.lang || 'ko';
+    this.threads = config.threads || 4;
+  }
+
+  async transcribe(audioPath, options = {}) {
+    return new Promise((resolve, reject) => {
+      const args = [
+        '-m', 'termux_stt.cli.main',
+        'transcribe',
+        '--engine', 'sherpa',
+        '--model', this.model,
+        '--lang', this.lang,
+        '--format', 'json',
+        audioPath
+      ];
+
+      const proc = spawn('python', args, { env: process.env });
+      let stdout = '';
+      let stderr = '';
+
+      proc.stdout.on('data', (d) => { stdout += d.toString(); });
+      proc.stderr.on('data', (d) => { stderr += d.toString(); });
+
+      proc.on('close', (code) => {
+        if (code !== 0) {
+          return reject(new Error(`Sherpa transcription failed (code ${code}): ${stderr}`));
+        }
+        try {
+          const parsed = JSON.parse(stdout);
+          const segments = (parsed.segments || []).map(
+            s => new Segment(s.start, s.end, s.text, s.speaker, s.confidence)
+          );
+          resolve(new TranscriptResult(parsed.text, segments, parsed.language || this.lang, parsed.duration));
+        } catch (e) {
+          resolve(new TranscriptResult(stdout.trim(), [new Segment(0, 0, stdout.trim())], this.lang));
+        }
+      });
+    });
+  }
+
+  async diarize(audioPath, numSpeakers = 2) {
+    return new Promise((resolve, reject) => {
+      const args = [
+        '-m', 'termux_stt.cli.main',
+        'diarize',
+        '--engine', 'sherpa',
+        '--speakers', String(numSpeakers),
+        '--format', 'json',
+        audioPath
+      ];
+
+      const proc = spawn('python', args, { env: process.env });
+      let stdout = '';
+      let stderr = '';
+
+      proc.stdout.on('data', (d) => { stdout += d.toString(); });
+      proc.stderr.on('data', (d) => { stderr += d.toString(); });
+
+      proc.on('close', (code) => {
+        if (code !== 0) {
+          return reject(new Error(`Sherpa diarization failed (code ${code}): ${stderr}`));
+        }
+        try {
+          const parsed = JSON.parse(stdout);
+          const segments = (parsed.segments || []).map(
+            s => new Segment(s.start, s.end, s.text, s.speaker, s.confidence)
+          );
+          resolve(new TranscriptResult(parsed.text, segments, parsed.language || this.lang, parsed.duration));
+        } catch (e) {
+          resolve(new TranscriptResult(stdout.trim(), [new Segment(0, 0, stdout.trim())], this.lang));
+        }
+      });
+    });
+  }
+
+  getInfo() {
+    return {
+      name: 'Sherpa-ONNX (Node.js)',
+      model: this.model,
+      language: this.lang,
+      threads: this.threads
+    };
+  }
+}
+
+module.exports = { SherpaEngine };

package/package.json

@@ -1,87 +1,87 @@
-{
-  "name": "termux-stt",
-  "version": "1.4.0",
-  "description": "On-device Speech-to-Text & speaker diarization framework utilizing device resources for Android Termux",
-  "main": "index.js",
-  "types": "index.d.ts",
-  "bin": {
-    "termux-stt": "bin/cli.js",
-    "termux-stt-install": "bin/termux-stt.js"
-  },
-  "files": [
-    "index.js",
-    "index.d.ts",
-    "lib",
-    "bin",
-    "README.md",
-    "LICENSE"
-  ],
-  "repository": {
-    "type": "git",
-    "url": "git+https://github.com/uno-km/termux-stt.git"
-  },
-  "keywords": [
-    "termux",
-    "stt",
-    "speech-to-text",
-    "whisper",
-    "whisper-cpp",
-    "vosk",
-    "sherpa-onnx",
-    "diarization",
-    "speaker-diarization",
-    "voice-recognition",
-    "audio-transcription",
-    "on-device-ai",
-    "edge-ai",
-    "mobile-ai",
-    "vulkan",
-    "vulkan-compute",
-    "gpu-acceleration",
-    "real-time-factor",
-    "low-latency",
-    "arm64",
-    "android",
-    "snapdragon",
-    "adreno",
-    "exynos",
-    "arm-mali",
-    "zero-compilation",
-    "vad",
-    "voice-activity-detection",
-    "silero-vad",
-    "x-vector",
-    "k-means",
-    "clustering",
-    "srt-export",
-    "vtt-export",
-    "rttm",
-    "microphone-streaming",
-    "headless-audio",
-    "pulseaudio",
-    "offline-speech",
-    "privacy-first",
-    "termux-aichain",
-    "termux-tts",
-    "termux-llamacpp",
-    "termux-diffusion",
-    "ameva-runtime",
-    "ggml",
-    "quantization",
-    "bionic-libc",
-    "autonomous-agents",
-    "voice-assistant"
-  ],
-  "author": "uno-km (AMEVA Foundation) <dev@ameva.org>",
-  "license": "MIT",
-  "bugs": {
-    "url": "https://github.com/uno-km/termux-stt/issues"
-  },
-  "homepage": "https://uno-km.github.io/termux-stt/",
-  "dependencies": {
-    "@ameva/runtime": ">=2.0.0"
-  },
-  "engines": {
-    "node": ">=16.0.0"
-  }
-}
+{
+  "name": "termux-stt",
+  "version": "2.0.0",
+  "description": "On-device Speech-to-Text & neural speaker diarization framework for Android Termux (Whisper Vulkan GPU & Sherpa-ONNX)",
+  "main": "index.js",
+  "types": "index.d.ts",
+  "bin": {
+    "termux-stt": "bin/cli.js",
+    "termux-stt-install": "bin/termux-stt.js"
+  },
+  "files": [
+    "index.js",
+    "index.d.ts",
+    "lib",
+    "bin",
+    "README.md",
+    "LICENSE"
+  ],
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/uno-km/termux-stt.git"
+  },
+  "keywords": [
+    "termux",
+    "stt",
+    "speech-to-text",
+    "whisper",
+    "whisper-cpp",
+    "vosk",
+    "sherpa-onnx",
+    "diarization",
+    "speaker-diarization",
+    "voice-recognition",
+    "audio-transcription",
+    "on-device-ai",
+    "edge-ai",
+    "mobile-ai",
+    "vulkan",
+    "vulkan-compute",
+    "gpu-acceleration",
+    "real-time-factor",
+    "low-latency",
+    "arm64",
+    "android",
+    "snapdragon",
+    "adreno",
+    "exynos",
+    "arm-mali",
+    "zero-compilation",
+    "vad",
+    "voice-activity-detection",
+    "silero-vad",
+    "x-vector",
+    "k-means",
+    "clustering",
+    "srt-export",
+    "vtt-export",
+    "rttm",
+    "microphone-streaming",
+    "headless-audio",
+    "pulseaudio",
+    "offline-speech",
+    "privacy-first",
+    "termux-aichain",
+    "termux-tts",
+    "termux-llamacpp",
+    "termux-diffusion",
+    "ameva-runtime",
+    "ggml",
+    "quantization",
+    "bionic-libc",
+    "autonomous-agents",
+    "voice-assistant"
+  ],
+  "author": "uno-km (AMEVA Foundation) <dev@ameva.org>",
+  "license": "MIT",
+  "bugs": {
+    "url": "https://github.com/uno-km/termux-stt/issues"
+  },
+  "homepage": "https://uno-km.github.io/termux-stt/",
+  "dependencies": {
+    "@ameva/runtime": ">=2.0.0"
+  },
+  "engines": {
+    "node": ">=16.0.0"
+  }
+}

package/lib/vosk.js

@@ -1,60 +0,0 @@
-/**
- * VoskEngine for Node.js
- */
-
-const { spawn } = require('child_process');
-const { Engine, TranscriptResult, Segment } = require('./engine');
-
-class VoskEngine extends Engine {
-  constructor(config = {}) {
-    super(config);
-    this.model = config.model || 'small-ko-0.22';
-    this.lang = config.lang || 'ko';
-  }
-
-  async transcribe(audioPath, options = {}) {
-    return new Promise((resolve, reject) => {
-      const args = [
-        '-m', 'termux_stt.cli.main',
-        'transcribe',
-        '--engine', 'vosk',
-        '--model', this.model,
-        '--lang', this.lang,
-        '--format', 'json',
-        audioPath
-      ];
-
-      const proc = spawn('python', args, { env: process.env });
-      let stdout = '';
-      let stderr = '';
-
-      proc.stdout.on('data', (d) => { stdout += d.toString(); });
-      proc.stderr.on('data', (d) => { stderr += d.toString(); });
-
-      proc.on('close', (code) => {
-        if (code !== 0) {
-          return reject(new Error(`Vosk transcription failed (code ${code}): ${stderr}`));
-        }
-        try {
-          const parsed = JSON.parse(stdout);
-          const segments = (parsed.segments || []).map(
-            s => new Segment(s.start, s.end, s.text, s.speaker, s.confidence)
-          );
-          resolve(new TranscriptResult(parsed.text, segments, parsed.language || this.lang, parsed.duration));
-        } catch (e) {
-          resolve(new TranscriptResult(stdout.trim(), [new Segment(0, 0, stdout.trim())], this.lang));
-        }
-      });
-    });
-  }
-
-  getInfo() {
-    return {
-      name: 'Vosk (Node.js)',
-      model: this.model,
-      language: this.lang
-    };
-  }
-}
-
-module.exports = { VoskEngine };