@qvac/sdk 0.8.1 → 1.1.0

package/dist/examples/llamacpp-native-tools.js

@@ -1,35 +1,45 @@
-import { z } from "zod";
-import { completion, loadModel, unloadModel, QWEN3_1_7B_INST_Q4, } from "../index.js";
-// Define Zod schemas for tool parameters
-const weatherSchema = z.object({
-    city: z.string().describe("City name"),
-    country: z.string().describe("Country code").optional(),
-});
-const horoscopeSchema = z.object({
-    sign: z.string().describe("An astrological sign like Taurus or Aquarius"),
-});
-// Map tool names to their schemas for runtime validation
-const toolSchemas = {
-    get_weather: weatherSchema,
-    get_horoscope: horoscopeSchema,
-};
-// Simple tool definitions - just name, description, and Zod schema!
+import { completion, loadModel, unloadModel, QWEN_3_1_7B_INST_Q4, } from "../index.js";
+// Tool definitions
 const tools = [
     {
+        type: "function",
         name: "get_weather",
         description: "Get current weather for a city",
-        parameters: weatherSchema,
+        parameters: {
+            type: "object",
+            properties: {
+                city: {
+                    type: "string",
+                    description: "City name",
+                },
+                country: {
+                    type: "string",
+                    description: "ISO country code",
+                },
+            },
+            required: ["city"],
+        },
     },
     {
+        type: "function",
         name: "get_horoscope",
         description: "Get today's horoscope for an astrological sign",
-        parameters: horoscopeSchema,
+        parameters: {
+            type: "object",
+            properties: {
+                sign: {
+                    type: "string",
+                    description: "An astrological sign like Taurus or Aquarius",
+                },
+            },
+            required: ["sign"],
+        },
     },
 ];
 try {
     // Load model from provided file path with tools support enabled
     const modelId = await loadModel({
-        modelSrc: QWEN3_1_7B_INST_Q4,
+        modelSrc: QWEN_3_1_7B_INST_Q4,
         modelType: "llm",
         modelConfig: {
             ctx_size: 4096,
@@ -78,67 +88,15 @@ try {
     if (toolCalls.length > 0) {
         for (const call of toolCalls) {
             console.log(`  - ${call.name}(${JSON.stringify(call.arguments)})`);
-            const schema = toolSchemas[call.name];
-            if (schema) {
-                const validated = schema.safeParse(call.arguments);
-                if (validated.success) {
-                    console.log(`    ✓ Arguments validated with Zod`);
-                }
-                else {
-                    console.log(`    ✗ Validation failed:`, validated.error);
-                }
-            }
         }
     }
     else {
         console.log("  No tool calls detected in response");
     }
     console.log("\n📊 Performance Stats:", stats);
-    // Execute tool calls and send results back to the model
-    if (toolCalls.length > 0) {
-        console.log("\n\n🔧 Simulating Tool Execution...");
-        // Simulate tool execution (in a real app, you'd call actual APIs)
-        const toolResults = toolCalls.map((call) => {
-            let result = "";
-            if (call.name === "get_weather") {
-                const args = call.arguments;
-                result = `The weather in ${args.city} is sunny, 22°C with light clouds.`;
-            }
-            else if (call.name === "get_horoscope") {
-                const args = call.arguments;
-                result = `Horoscope for ${args.sign}: Today is a great day for new beginnings and creative endeavors!`;
-            }
-            console.log(`  ✓ ${call.name}: ${result}`);
-            return { toolCallId: call.id, result };
-        });
-        // Add tool results to conversation history
-        history.push({
-            role: "assistant",
-            content: await result.text,
-        });
-        // Add tool results as tool messages
-        for (const toolResult of toolResults) {
-            history.push({
-                role: "tool",
-                content: toolResult.result,
-            });
-        }
-        // Send follow-up question with tool results
-        console.log("\n\n🤖 Follow-up Response with Tool Results:");
-        const followUpResult = completion({
-            modelId,
-            history,
-            stream: true,
-            tools,
-        });
-        for await (const token of followUpResult.tokenStream) {
-            process.stdout.write(token);
-        }
-        const followUpStats = await followUpResult.stats;
-        console.log("\n\n📊 Follow-up Stats:", followUpStats);
-    }
     console.log("\n\n🎉 Completed!");
     await unloadModel({ modelId, clearStorage: false });
+    process.exit(0);
 }
 catch (error) {
     console.error("❌ Error:", error);

package/dist/examples/multi-model-demo.js

@@ -1,7 +1,7 @@
-import { completion, LLAMA_3_2_1B_INST_Q4_0, loadModel, QWEN3_1_7B_INST_Q4, unloadModel, } from "@qvac/sdk";
+import { completion, LLAMA_3_2_1B_INST_Q4_0, loadModel, QWEN_3_1_7B_INST_Q4, unloadModel, } from "@qvac/sdk";
 try {
     const alice = await loadModel({
-        modelSrc: QWEN3_1_7B_INST_Q4,
+        modelSrc: QWEN_3_1_7B_INST_Q4,
         modelType: "llm",
         modelConfig: {
             ctx_size: 4096,
@@ -67,6 +67,7 @@ try {
     console.log("📊 Bob Stats: ", await bobResponse2.stats);
     await unloadModel({ modelId: alice, clearStorage: false });
     await unloadModel({ modelId: bob, clearStorage: false });
+    process.exit(0);
 }
 catch (error) {
     console.error("❌ Error:", error);

package/dist/examples/parallel-download.js

@@ -1,51 +1,19 @@
-import { LLAMA_3_2_1B_INST_Q4_0, WHISPER_TINY, VAD_SILERO_5_1_2, GTE_LARGE_FP16, downloadAsset, close, } from "@qvac/sdk";
-function now() {
-    return Date.now();
-}
-const assets = [
-    { name: "Llama 3.2 1B", src: LLAMA_3_2_1B_INST_Q4_0 },
-    { name: "Whisper Tiny", src: WHISPER_TINY },
-    { name: "VAD Silero", src: VAD_SILERO_5_1_2 },
-    { name: "GTE Large FP16", src: GTE_LARGE_FP16 },
-];
-const timers = {};
-console.log(`\n=== Parallel Download (${assets.length} assets) ===\n`);
-const wallStart = now();
+import { LLAMA_3_2_1B_INST_Q4_0, downloadAsset, WHISPER_LARGE_3, } from "@qvac/sdk";
 try {
-    const promises = assets.map((asset) => {
-        timers[asset.name] = { start: now() };
-        return downloadAsset({
-            assetSrc: asset.src,
-            onProgress: (progress) => {
-                const t = timers[asset.name];
-                if (t.firstProgress == null) {
-                    t.firstProgress = now();
-                    console.log(`[${asset.name}] first progress at ${((t.firstProgress - wallStart) / 1000).toFixed(1)}s — ${progress.percentage}%`);
-                }
-                if (progress.percentage === 100 && t.end == null) {
-                    t.end = now();
-                    console.log(`[${asset.name}] done at ${((t.end - wallStart) / 1000).toFixed(1)}s`);
-                }
-            },
-        });
+    const llamaDownloadPromise = downloadAsset({
+        assetSrc: LLAMA_3_2_1B_INST_Q4_0,
+        onProgress: (progress) => {
+            console.log("💬 Llama progress:", progress);
+        },
+    });
+    const whisperDownloadPromise = downloadAsset({
+        assetSrc: WHISPER_LARGE_3,
+        onProgress: (progress) => {
+            console.log("🔊 Whisper progress:", progress);
+        },
     });
-    const results = await Promise.allSettled(promises);
-    const wallEnd = now();
-    console.log(`\n=== Results ===\n`);
-    for (let i = 0; i < assets.length; i++) {
-        const asset = assets[i];
-        const result = results[i];
-        const t = timers[asset.name];
-        const status = result.status === "fulfilled" ? "OK" : "FAILED";
-        const reason = result.status === "rejected" ? ` — ${result.reason}` : "";
-        const timeToFirst = t.firstProgress != null
-            ? `${((t.firstProgress - t.start) / 1000).toFixed(1)}s`
-            : "N/A";
-        const total = t.end != null ? `${((t.end - t.start) / 1000).toFixed(1)}s` : "N/A";
-        console.log(`${status} ${asset.name}: first-progress=${timeToFirst}, total=${total}${reason}`);
-    }
-    console.log(`\nWall time: ${((wallEnd - wallStart) / 1000).toFixed(1)}s`);
-    void close();
+    await Promise.all([llamaDownloadPromise, whisperDownloadPromise]);
+    process.exit(0);
 }
 catch (error) {
     console.error("❌ Error:", error);

package/dist/examples/rag/rag-chromadb.js

@@ -122,6 +122,7 @@ try {
         console.log();
     }
     await unloadModel({ modelId });
+    process.exit(0);
 }
 catch (error) {
     console.error("❌ Error:", error);

package/dist/examples/rag/rag-hyperdb-workspaces.d.ts

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=rag-hyperdb-workspaces.d.ts.map

package/dist/examples/rag/rag-hyperdb-workspaces.js

@@ -0,0 +1,107 @@
+import { loadModel, unloadModel, GTE_LARGE_FP16, ragSaveEmbeddings, ragSearch, ragDeleteEmbeddings, } from "@qvac/sdk";
+try {
+    console.log("🚀 RAG Workspaces with Chunking Example\n");
+    // Load model
+    const modelId = await loadModel({
+        modelSrc: GTE_LARGE_FP16,
+        modelType: "embeddings",
+        onProgress: (p) => process.stdout.write(`\rLoading: ${Math.round(p.percentage * 100)}%`),
+    });
+    console.log("\n");
+    // Medical article (will be chunked)
+    const medicalArticle = `
+    COVID-19: A Comprehensive Overview
+    
+    COVID-19 is a respiratory illness caused by the SARS-CoV-2 virus, first identified in December 2019 
+    in Wuhan, China. The virus spreads primarily through respiratory droplets when an infected person 
+    coughs, sneezes, or talks.
+    
+    Common symptoms include fever, cough, fatigue, and loss of taste or smell. Some patients may 
+    experience shortness of breath, muscle aches, sore throat, and headaches. Symptoms typically 
+    appear 2-14 days after exposure to the virus.
+    
+    Vaccines have been developed using various technologies including mRNA, viral vector, and 
+    inactivated virus approaches. These vaccines have proven effective in preventing severe illness, 
+    hospitalization, and death. Booster doses are recommended to maintain immunity.
+    
+    Treatment options vary based on severity. Mild cases often require only supportive care including 
+    rest and hydration. Severe cases may require hospitalization, oxygen therapy, and antiviral 
+    medications such as remdesivir or Paxlovid.
+  `;
+    // Tech documents (shorter, won't be chunked)
+    const techDocs = [
+        "Artificial intelligence is transforming industries through automation and advanced data analysis capabilities.",
+        "Deep learning neural networks enable breakthroughs in computer vision and natural language processing.",
+    ];
+    // Save medical article with chunking
+    console.log("📚 Saving medical article with chunking...");
+    const medicalResult = await ragSaveEmbeddings({
+        modelId,
+        workspace: "medical",
+        documents: medicalArticle,
+        chunk: true,
+        chunkOpts: {
+            chunkSize: 100,
+            chunkOverlap: 10,
+            chunkStrategy: "paragraph",
+            splitStrategy: "token",
+        },
+    });
+    console.log(`✅ Created ${medicalResult.processed.length} chunks from medical article`);
+    // Save tech docs without chunking
+    console.log("\n📚 Saving tech documents...");
+    const techResult = await ragSaveEmbeddings({
+        modelId,
+        workspace: "technology",
+        documents: techDocs,
+        chunk: false,
+    });
+    console.log(`✅ Saved ${techResult.processed.length} tech documents`);
+    // Test searches
+    const searches = [
+        { workspace: "medical", query: "COVID symptoms fever", label: "Medical" },
+        { workspace: "medical", query: "vaccine technology", label: "Medical" },
+        { workspace: "technology", query: "neural networks", label: "Tech" },
+        { workspace: "technology", query: "COVID", label: "Tech (isolation test)" },
+    ];
+    console.log("\n🔍 Running searches:");
+    for (const { workspace, query, label } of searches) {
+        const results = await ragSearch({ modelId, workspace, query, topK: 1 });
+        console.log(`\n${label}: "${query}"`);
+        if (results.length > 0 && results[0]) {
+            console.log(`  ✓ Score: ${results[0].score.toFixed(3)}`);
+            console.log(`  ✓ Match: ${results[0].content.substring(0, 80)}...`);
+        }
+        else {
+            console.log(`  ✗ No results (workspace isolation working correctly)`);
+        }
+    }
+    // Test default workspace
+    console.log("\n📁 Testing default workspace...");
+    await ragSaveEmbeddings({
+        modelId,
+        documents: ["Default workspace content for testing"],
+        chunk: false,
+    });
+    const defaultSearch = await ragSearch({
+        modelId,
+        query: "default workspace",
+    });
+    console.log(`Default workspace: ${defaultSearch?.[0]?.content === "Default workspace content for testing" ? "✅ Working" : "❌ Failed"}`);
+    // Cleanup example
+    const firstChunk = medicalResult.processed.find((p) => p.status === "fulfilled" && p.id);
+    if (firstChunk?.id) {
+        await ragDeleteEmbeddings({
+            modelId,
+            workspace: "medical",
+            ids: [firstChunk.id],
+        });
+        console.log("\n🗑️  Deleted one medical chunk");
+    }
+    await unloadModel({ modelId });
+    process.exit(0);
+}
+catch (error) {
+    console.error("❌ Error:", error);
+    process.exit(1);
+}

package/dist/examples/rag/rag-hyperdb.d.ts

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=rag-hyperdb.d.ts.map

package/dist/examples/rag/{rag-hyperdb/ingest.js → rag-hyperdb.js}

@@ -1,8 +1,7 @@
-import { loadModel, unloadModel, GTE_LARGE_FP16, ragIngest, ragSearch, ragCloseWorkspace, } from "@qvac/sdk";
+import { loadModel, unloadModel, GTE_LARGE_FP16, ragSaveEmbeddings, ragSearch, } from "@qvac/sdk";
 try {
     // Get query from command line or use default
     const query = process.argv[2] || "machine learning algorithms";
-    const workspace = "ingest-example";
     console.log(`🔍 Query: "${query}"`);
     const modelId = await loadModel({
         modelSrc: GTE_LARGE_FP16,
@@ -21,18 +20,16 @@ try {
         "Cloud computing delivers computing services over the internet, allowing users to access resources like storage, processing power, and applications on-demand from anywhere.",
         "Cybersecurity protects digital systems, networks, and data from malicious attacks, unauthorized access, and various forms of cyber threats through multiple layers of defense.",
     ];
-    console.log("📚 Ingesting documents...");
-    const result = await ragIngest({
+    console.log("📚 Embedding documents...");
+    const docs = await ragSaveEmbeddings({
         modelId,
-        workspace,
         documents: samples,
         chunk: false,
     });
-    console.log(`✅ Ingested ${result.processed.length} documents`);
+    console.log("🔍 Documents:", docs);
     console.log("🔎 Searching for similar documents...");
     const results = await ragSearch({
         modelId,
-        workspace,
         query,
         topK: 3,
     });
@@ -44,10 +41,8 @@ try {
         console.log("=".repeat(100));
         console.log();
     });
-    // Cleanup: close and delete workspace
-    await ragCloseWorkspace({ workspace, deleteOnClose: true });
-    console.log(`\n🗑️  Deleted '${workspace}' workspace`);
     await unloadModel({ modelId });
+    process.exit(0);
 }
 catch (error) {
     console.error("❌ Error:", error);

package/dist/examples/rag/rag-lancedb.js

@@ -87,6 +87,7 @@ try {
         console.log();
     });
     await unloadModel({ modelId });
+    process.exit(0);
 }
 catch (error) {
     console.error("❌ Error:", error);

package/dist/examples/rag/rag-sqlite.js

@@ -99,6 +99,7 @@ try {
     });
     await unloadModel({ modelId });
     db.close();
+    process.exit(0);
 }
 catch (error) {
     console.error("❌ Error:", error);

package/dist/examples/seed-hyperdrive.d.ts

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=seed-hyperdrive.d.ts.map

package/dist/examples/text-to-speech.d.ts

@@ -0,0 +1,3 @@
+export declare function createWav(audioBuffer: number[], sampleRate?: number, filename?: string): void;
+export declare function playAudio(audioBuffer: Buffer): void;
+//# sourceMappingURL=text-to-speech.d.ts.map

package/dist/examples/{tts/utils.js → text-to-speech.js}

@@ -1,10 +1,61 @@
+import { loadModel, textToSpeech, unloadModel, TTS_PIPER_NORMAN_EN_US_ONNX_MEDIUM, TTS_PIPER_NORMAN_EN_US_ONNX_MEDIUM_CONFIG, } from "@qvac/sdk";
 import { writeFileSync, unlinkSync } from "fs";
 import { spawnSync } from "child_process";
 import { platform } from "os";
-/**
- * Create WAV header for 16-bit PCM audio
- */
-export function createWavHeader(dataLength, sampleRate) {
+// Get eSpeakDataPath from command line arguments or use default
+const eSpeakDataPath = process.argv[2];
+console.log(`Using eSpeak data path: ${eSpeakDataPath}`);
+if (!eSpeakDataPath) {
+    console.error("eSpeakDataPath is required");
+    process.exit(1);
+}
+try {
+    const modelId = await loadModel({
+        modelSrc: TTS_PIPER_NORMAN_EN_US_ONNX_MEDIUM,
+        modelType: "tts",
+        configSrc: TTS_PIPER_NORMAN_EN_US_ONNX_MEDIUM_CONFIG,
+        eSpeakDataPath,
+        modelConfig: {
+            language: "en",
+        },
+        onProgress: (progress) => {
+            console.log(progress);
+        },
+    });
+    console.log(`Model loaded: ${modelId}`);
+    // Test TTS
+    console.log("🎵 Testing Text-to-Speech...");
+    const result = textToSpeech({
+        modelId,
+        text: "Hello, how are you today? This is a test of the TTS system.",
+        inputType: "text",
+        stream: false,
+    });
+    const audioBuffer = await result.buffer;
+    console.log(`TTS complete. Total bytes: ${audioBuffer.length}`);
+    // Save and play audio
+    console.log("💾 Saving audio to file...");
+    createWav(audioBuffer, 22050, "tts-output.wav");
+    console.log("✅ Audio saved to tts-output.wav");
+    console.log("🔊 Playing audio...");
+    const audioData = int16ArrayToBuffer(audioBuffer);
+    const wavBuffer = Buffer.concat([
+        createWavHeader(audioData.length),
+        audioData,
+    ]);
+    playAudio(wavBuffer);
+    console.log("✅ Audio playback complete");
+    // Unload the model
+    await unloadModel({ modelId });
+    console.log("Model unloaded");
+    process.exit(0);
+}
+catch (error) {
+    console.error("❌ Error:", error);
+    process.exit(1);
+}
+// Function to create WAV header for 16-bit PCM audio
+function createWavHeader(dataLength, sampleRate = 22050) {
     const header = Buffer.alloc(44);
     // RIFF header
     header.write("RIFF", 0);
@@ -24,30 +75,24 @@ export function createWavHeader(dataLength, sampleRate) {
     header.writeUInt32LE(dataLength, 40);
     return header;
 }
-/**
- * Convert Int16Array to Buffer
- */
-export function int16ArrayToBuffer(samples) {
-    const buffer = Buffer.alloc(samples.length * 2);
-    for (let i = 0; i < samples.length; i++) {
-        const value = Math.max(-32768, Math.min(32767, Math.round(samples[i] ?? 0)));
+// Function to convert Int16Array to Buffer
+function int16ArrayToBuffer(int16Array) {
+    const buffer = Buffer.alloc(int16Array.length * 2);
+    for (let i = 0; i < int16Array.length; i++) {
+        const value = int16Array[i] ?? 0;
         buffer.writeInt16LE(value, i * 2);
     }
     return buffer;
 }
-/**
- * Create and save WAV file
- */
-export function createWav(audioBuffer, sampleRate, filename) {
+// Function to create and save WAV file (similar to reference implementation)
+export function createWav(audioBuffer, sampleRate = 22050, filename = "output.wav") {
     const audioData = int16ArrayToBuffer(audioBuffer);
     const wavHeader = createWavHeader(audioData.length, sampleRate);
     const wavFile = Buffer.concat([wavHeader, audioData]);
     writeFileSync(filename, wavFile);
     console.log(`WAV file saved as: ${filename}`);
 }
-/**
- * Play audio using system audio players
- */
+// Function to play audio using system audio players
 export function playAudio(audioBuffer) {
     const currentPlatform = platform();
     const tempFile = `/tmp/audio-${Date.now()}.wav`;

package/dist/examples/translation/translation-indic.js

@@ -1,7 +1,4 @@
 import { loadModel, translate, unloadModel, MARIAN_EN_HI_INDIC_1B_Q4_0, } from "@qvac/sdk";
-// NOTE: @qvac/translation-nmtcpp version 0.1.6 does not work well with IndicTrans models
-// This example is expected to fail
-// A fix is in progress for future releases
 try {
     const modelId = await loadModel({
         modelSrc: MARIAN_EN_HI_INDIC_1B_Q4_0,
@@ -10,7 +7,6 @@ try {
             console.log(progress);
         },
         modelConfig: {
-            engine: "IndicTrans",
             from: "eng_Latn",
             to: "hin_Deva",
         },
@@ -26,6 +22,7 @@ try {
     const translatedText = await result.text;
     console.log(`Translated text EN -> HI: ${text} -> "${translatedText}"`);
     await unloadModel({ modelId });
+    process.exit(0);
 }
 catch (error) {
     console.error("❌ Error:", error);

package/dist/examples/translation/translation-llm-context.js

@@ -22,6 +22,7 @@ try {
     const translatedTextExplicit = await resultExplicit.text;
     console.log(`${engText} -> "${translatedTextExplicit}"`); // "banco" (not "orilla")
     await unloadModel({ modelId, clearStorage: false });
+    process.exit(0);
 }
 catch (error) {
     console.error("❌ Error:", error);

package/dist/examples/translation/translation-llm.js

@@ -1,4 +1,4 @@
-import { translate, loadModel, unloadModel, SALAMANDRATA_2B_INST_Q4, } from "@qvac/sdk";
+import { translate, loadModel, SALAMANDRATA_2B_INST_Q4 } from "@qvac/sdk";
 try {
     const modelId = await loadModel({
         modelSrc: SALAMANDRATA_2B_INST_Q4,
@@ -30,7 +30,7 @@ try {
     const translatedTextAutodetect = await resultAutodetect.text;
     console.log(`Explicit source: ${engText} -> "${translatedTextExplicit}"`);
     console.log(`Autodetected source: ${espText} -> "${translatedTextAutodetect}"`);
-    await unloadModel({ modelId, clearStorage: false });
+    process.exit(0);
 }
 catch (error) {
     console.error("❌ Error:", error);

package/dist/examples/translation/translation-opus.js

@@ -1,15 +1,11 @@
-import { loadModel, translate, unloadModel, MARIAN_OPUS_EN_IT_Q4_0, } from "@qvac/sdk";
+import { loadModel, translate, unloadModel, MARIAN_OPUS_EN_IT_Q0F32, } from "@qvac/sdk";
 try {
     const modelId = await loadModel({
-        modelSrc: MARIAN_OPUS_EN_IT_Q4_0,
+        modelSrc: MARIAN_OPUS_EN_IT_Q0F32,
         modelType: "nmt",
         modelConfig: {
-            engine: "Opus",
             from: "en",
             to: "it",
-            temperature: 0.2,
-            norepeatngramsize: 3,
-            lengthpenalty: 1.2,
         },
         onProgress: (progress) => {
             console.log(progress);
@@ -26,6 +22,7 @@ try {
     const translatedText = await result.text;
     console.log(`Translated text EN -> IT: ${text} -> "${translatedText}"`);
     await unloadModel({ modelId });
+    process.exit(0);
 }
 catch (error) {
     console.error("❌ Error:", error);

package/dist/examples/translation/translation-stream.js

@@ -1,10 +1,9 @@
-import { loadModel, translate, unloadModel, MARIAN_OPUS_EN_IT_Q4_0, } from "@qvac/sdk";
+import { loadModel, translate, MARIAN_OPUS_EN_IT_Q0F32, unloadModel, } from "@qvac/sdk";
 try {
     const modelId = await loadModel({
-        modelSrc: MARIAN_OPUS_EN_IT_Q4_0,
+        modelSrc: MARIAN_OPUS_EN_IT_Q0F32,
         modelType: "nmt",
         modelConfig: {
-            engine: "Opus",
             from: "en",
             to: "it",
         },
@@ -28,6 +27,7 @@ try {
         console.log(`Processing stats:`, stats);
     }
     await unloadModel({ modelId, clearStorage: false });
+    process.exit(0);
 }
 catch (error) {
     console.error("❌ Error:", error);

package/dist/examples/{transcription/whispercpp-filesystem.js → whispercpp-filesystem.js}

@@ -1,4 +1,4 @@
-import { loadModel, unloadModel, transcribe, WHISPER_TINY } from "@qvac/sdk";
+import { loadModel, unloadModel, transcribe, WHISPER_TINY, VAD_SILERO_5_1_2, } from "@qvac/sdk";
 // Parse command line arguments
 const args = process.argv.slice(2);
 if (!args[0]) {
@@ -8,12 +8,16 @@ if (!args[0]) {
 const audioFilePath = args[0];
 try {
     console.log("🎤 Starting Whisper transcription example...");
-    // Load the Whisper model
+    // Load the Whisper model with VAD
     console.log("📥 Loading Whisper model...");
     const modelId = await loadModel({
         modelSrc: WHISPER_TINY,
         modelType: "whisper",
+        vadModelSrc: VAD_SILERO_5_1_2,
         modelConfig: {
+            mode: "caption",
+            output_format: "plaintext",
+            min_seconds: 2,
             audio_format: "f32le",
             // Sampling strategy
             strategy: "greedy",
@@ -62,6 +66,7 @@ try {
     console.log("🧹 Unloading Whisper model...");
     await unloadModel({ modelId });
     console.log("✅ Whisper model unloaded successfully");
+    process.exit(0);
 }
 catch (error) {
     console.error("❌ Error:", error);