diff --git a/examples/audio-demo/package.json b/examples/audio-demo/package.json
index dcbc35d..2dd8082 100644
--- a/examples/audio-demo/package.json
+++ b/examples/audio-demo/package.json
@@ -12,7 +12,7 @@
     "deploy": "gh-pages -d dist"
   },
   "dependencies": {
-    "@browserai/browserai": "^1.0.13",
+    "@browserai/browserai": "^1.0.16",
     "@emotion/styled": "^11.14.0",
     "@huggingface/transformers": "^3.2.4",
     "@mlc-ai/web-llm": "^0.2.77",
diff --git a/examples/audio-demo/src/components/ChatInterface.tsx b/examples/audio-demo/src/components/ChatInterface.tsx
index d8eb1c8..7d2e07a 100644
--- a/examples/audio-demo/src/components/ChatInterface.tsx
+++ b/examples/audio-demo/src/components/ChatInterface.tsx
@@ -719,7 +719,9 @@ export default function ChatInterface() {
       const audioLoadStart = performance.now();
       setStatus('Loading audio model...');
       try {
-        await audioAIRef.loadModel('whisper-tiny-en');
+        await audioAIRef.loadModel('whisper-tiny-en', {
+          device: 'webgpu'
+        });
         console.log('Audio model loaded successfully');
       } catch (error) {
         console.error('Error loading audio model:', error);
@@ -733,7 +735,7 @@ export default function ChatInterface() {
       const ttsLoadStart = performance.now();
       setStatus('Loading TTS model...');
       try {
-        await ttsAIRef.loadModel('speecht5-tts');
+        await ttsAIRef.loadModel('kokoro-tts');
         console.log('TTS model loaded successfully');
       } catch (error) {
         console.error('Error loading TTS model:', error);
diff --git a/examples/tts-demo/package.json b/examples/tts-demo/package.json
index d4d2c8e..bc29598 100644
--- a/examples/tts-demo/package.json
+++ b/examples/tts-demo/package.json
@@ -10,7 +10,7 @@
     "preview": "vite preview"
   },
   "dependencies": {
-    "@browserai/browserai": "^1.0.15",
+    "@browserai/browserai": "^1.0.16",
     "@emotion/styled": "^11.14.0",
     "react": "^18.3.1",
     "react-dom": "^18.3.1"
diff --git a/package.json b/package.json
index a1e6538..402817f 100644
--- a/package.json
+++ b/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@browserai/browserai",
-  "version": "1.0.15",
+  "version": "1.0.16",
   "private": false,
   "description": "A library for running AI models directly in the browser",
   "main": "dist/index.js",
diff --git a/src/libs/transformers/backends/onnx.ts b/src/libs/transformers/backends/onnx.ts
index 1580401..43e26b7 100644
--- a/src/libs/transformers/backends/onnx.ts
+++ b/src/libs/transformers/backends/onnx.ts
@@ -109,7 +109,6 @@ export async function createInferenceSession(
   session_options: any,
   session_config: any
 ) {
-  const preferredBackend = session_config.backend || 'webgpu';
   if (wasmInitPromise) {
     // A previous session has already initialized the WASM runtime
     // so we wait for it to resolve before creating this new session.
@@ -142,9 +141,13 @@ if (ONNX_ENV?.wasm) {
   // https://onnxruntime.ai/docs/api/js/interfaces/Env.WebAssemblyFlags.html#wasmPaths
   // We use remote wasm files by default to make it easier for newer users.
   // In practice, users should probably self-host the necessary .wasm files.
-  ONNX_ENV.wasm.wasmPaths = `https://cdn.jsdelivr.net/npm/@huggingface/transformers@${env.version}/dist/`;
+
+  ONNX_ENV.wasm.wasmPaths = {
+    wasm: `https://cdn.jsdelivr.net/npm/@huggingface/transformers@${env.version}/dist/ort-wasm-simd-threaded.jsep.wasm`,
+    mjs: `https://cdn.jsdelivr.net/npm/@huggingface/transformers@${env.version}/dist/ort-wasm-simd-threaded.jsep.mjs`,
+  };
   // https://cdn.jsdelivr.net/npm/@huggingface/transformers@3.2.4/dist/ort-wasm-simd-threaded.jsep.wasm
-  // console.log('ONNX_ENV.wasm.wasmPaths', ONNX_ENV.wasm.wasmPaths);
+  console.log('ONNX_ENV.wasm.wasmPaths', ONNX_ENV.wasm.wasmPaths);
   // TODO: Add support for loading WASM files from cached buffer when we upgrade to onnxruntime-web@1.19.0
   // https://github.com/microsoft/onnxruntime/pull/21534
 
@@ -152,6 +155,9 @@ if (ONNX_ENV?.wasm) {
   // However, this is not necessary when using WebGPU, so we default to false.
   ONNX_ENV.wasm.proxy = false;
 
+  // Disable threading to avoid potential issues
+  ONNX_ENV.wasm.numThreads = 1;
+  
   // https://developer.mozilla.org/en-US/docs/Web/API/crossOriginIsolated
   if (typeof crossOriginIsolated === 'undefined' || !crossOriginIsolated) {
     ONNX_ENV.wasm.numThreads = 1;
@@ -159,6 +165,7 @@ if (ONNX_ENV?.wasm) {
 }
 
 if (ONNX_ENV?.webgpu) {
+  console.log('ONNX_ENV.webgpu', ONNX_ENV.webgpu);
   ONNX_ENV.webgpu.powerPreference = 'high-performance';
 }
 
diff --git a/src/libs/transformers/backends/ort-wasm-simd-threaded.jsep.wasm b/src/libs/transformers/backends/ort-wasm-simd-threaded.jsep.wasm
new file mode 100644
index 0000000..da2b62a
Binary files /dev/null and b/src/libs/transformers/backends/ort-wasm-simd-threaded.jsep.wasm differ
diff --git a/src/libs/transformers/env.ts b/src/libs/transformers/env.ts
index be0e4b4..17cf9b2 100644
--- a/src/libs/transformers/env.ts
+++ b/src/libs/transformers/env.ts
@@ -25,7 +25,7 @@
 import path from 'path';
 import url from 'url';
 
-const VERSION = '3.2.4';
+const VERSION = '3.3.1';
 
 // Check if various APIs are available (depends on environment)
 const IS_BROWSER_ENV = typeof window !== 'undefined' && typeof window.document !== 'undefined';
diff --git a/vite.config.js b/vite.config.js
new file mode 100644
index 0000000..0519ecb
--- /dev/null
+++ b/vite.config.js
@@ -0,0 +1 @@
+ 
\ No newline at end of file