Fix speech-to-text with Groq Whisper

certs/localhost.crt

@@ -0,0 +1,17 @@
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----

package-lock.json

@@ -16,7 +16,8 @@
         "google-tts-api": "^2.0.1"
       },
       "devDependencies": {
-        "puppeteer": "^25.3.0"
+        "puppeteer": "^25.3.0",
+        "selfsigned": "^2.1.0"
       }
     },
     "node_modules/@puppeteer/browsers": {
@@ -48,6 +49,26 @@
         }
       }
     },
+    "node_modules/@types/node": {
+      "version": "26.1.0",
+      "resolved": "https://registry.npmjs.org/@types/node/-/node-26.1.0.tgz",
+      "integrity": "sha512-O0A1G3xPGy4w7AgQdAQYUlQ+BKk2Oovw8eRpofyp5KdBZULnbe+WqaOVNrm705SHphCiG4XHsACrSmPu1f+Kgw==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "undici-types": "~8.3.0"
+      }
+    },
+    "node_modules/@types/node-forge": {
+      "version": "1.3.14",
+      "resolved": "https://registry.npmjs.org/@types/node-forge/-/node-forge-1.3.14.tgz",
+      "integrity": "sha512-mhVF2BnD4BO+jtOp7z1CdzaK4mbuK0LLQYAvdOLqHTavxFNq4zA1EmYkpnFjP8HOUzedfQkRnp0E2ulSAYSzAw==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@types/node": "*"
+      }
+    },
     "node_modules/accepts": {
       "version": "1.3.8",
       "resolved": "https://registry.npmjs.org/accepts/-/accepts-1.3.8.tgz",
@@ -920,6 +941,16 @@
         "node": ">= 0.6"
       }
     },
+    "node_modules/node-forge": {
+      "version": "1.4.0",
+      "resolved": "https://registry.npmjs.org/node-forge/-/node-forge-1.4.0.tgz",
+      "integrity": "sha512-LarFH0+6VfriEhqMMcLX2F7SwSXeWwnEAJEsYm5QKWchiVYVvJyV9v7UDvUv+w5HO23ZpQTXDv/GxdDdMyOuoQ==",
+      "dev": true,
+      "license": "(BSD-3-Clause OR GPL-2.0)",
+      "engines": {
+        "node": ">= 6.13.0"
+      }
+    },
     "node_modules/object-assign": {
       "version": "4.1.1",
       "resolved": "https://registry.npmjs.org/object-assign/-/object-assign-4.1.1.tgz",
@@ -1096,6 +1127,20 @@
       "integrity": "sha512-YZo3K82SD7Riyi0E1EQPojLz7kpepnSQI9IyPbHHg1XXXevb5dJI7tpyN2ADxGcQbHG7vcyRHk0cbwqcQriUtg==",
       "license": "MIT"
     },
+    "node_modules/selfsigned": {
+      "version": "2.4.1",
+      "resolved": "https://registry.npmjs.org/selfsigned/-/selfsigned-2.4.1.tgz",
+      "integrity": "sha512-th5B4L2U+eGLq1TVh7zNRGBapioSORUeymIydxgFpwww9d2qyKvtuPU2jJuHvYAwwqi2Y596QBL3eEqcPEYL8Q==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@types/node-forge": "^1.3.0",
+        "node-forge": "^1"
+      },
+      "engines": {
+        "node": ">=10"
+      }
+    },
     "node_modules/send": {
       "version": "0.19.2",
       "resolved": "https://registry.npmjs.org/send/-/send-0.19.2.tgz",
@@ -1291,6 +1336,13 @@
       "dev": true,
       "license": "MIT"
     },
+    "node_modules/undici-types": {
+      "version": "8.3.0",
+      "resolved": "https://registry.npmjs.org/undici-types/-/undici-types-8.3.0.tgz",
+      "integrity": "sha512-j375ScV60dom+YkPFIfTLcOiPxkN/buHz5GobjLhixFuANaNs3C9l4GmrWqejgXWJ7BbJcFYpTEUkS1Ge8bpZQ==",
+      "dev": true,
+      "license": "MIT"
+    },
     "node_modules/unpipe": {
       "version": "1.0.0",
       "resolved": "https://registry.npmjs.org/unpipe/-/unpipe-1.0.0.tgz",

package.json

@@ -4,6 +4,8 @@
   "main": "server.js",
   "scripts": {
     "start": "node server.js",
+    "start:https": "SSL_KEY_PATH=certs/localhost.key SSL_CERT_PATH=certs/localhost.crt node server.js",
+    "generate-cert": "node scripts/generate-cert.js",
     "check:permissions": "node scripts/permission-check.js"
   },
   "dependencies": {

script.js

@@ -411,12 +411,23 @@ async function transcribeAudio(blob) {
             body: blob
         });
 
+        const data = await response.json().catch(() => null);
         if (!response.ok) {
-            throw new Error('Transcription failed');
+            const errorMessage = data?.error || 'Voice transcription failed';
+            throw new Error(errorMessage);
+        }
+
+        const transcript = data?.transcript?.trim() || '';
+        if (!transcript) {
+            sourceText.value = 'No speech detected';
+            if (charCount) {
+                charCount.textContent = 0;
+            }
+            showStatus('No speech detected', 'error');
+            voiceStatus.textContent = '⚠️ No speech detected';
+            return;
         }
 
-        const data = await response.json();
-        const transcript = data.transcript || '';
         sourceText.value = transcript;
         if (charCount) {
             charCount.textContent = transcript.length;
@@ -428,8 +439,9 @@ async function transcribeAudio(blob) {
         await translateText();
     } catch (error) {
         console.error('Voice transcription error:', error);
-        showStatus('Voice transcription failed', 'error');
+        const errorMessage = error?.message || 'Voice transcription failed';
-        voiceStatus.textContent = '❌ Voice transcription failed';
+        showStatus(errorMessage, 'error');
+        voiceStatus.textContent = `❌ ${errorMessage}`;
     }
 }
 

scripts/generate-cert.js

@@ -0,0 +1,21 @@
+const fs = require('fs');
+const path = require('path');
+const selfsigned = require('selfsigned');
+
+const outDir = path.resolve(__dirname, '..', 'certs');
+fs.mkdirSync(outDir, { recursive: true });
+
+const attrs = [{ name: 'commonName', value: 'localhost' }];
+const opts = { days: 365, keySize: 2048, extensions: [{ name: 'subjectAltName', altNames: [{ type: 2, value: 'localhost' }, { type: 7, ip: '127.0.0.1' }] }] };
+const pems = selfsigned.generate(attrs, opts);
+
+const keyPath = path.join(outDir, 'localhost.key');
+const certPath = path.join(outDir, 'localhost.crt');
+
+fs.writeFileSync(keyPath, pems.private, { mode: 0o600 });
+fs.writeFileSync(certPath, pems.cert, { mode: 0o644 });
+
+console.log('Wrote key:', keyPath);
+console.log('Wrote cert:', certPath);
+console.log('To start HTTPS server using these certs run:');
+console.log('  SSL_KEY_PATH=' + keyPath + ' SSL_CERT_PATH=' + certPath + ' npm start');

server.js

@@ -7,6 +7,7 @@ const googleTTS = require('google-tts-api');
 require('dotenv').config();
 
 const HUGGINGFACE_API_TOKEN = process.env.HUGGINGFACEHUB_API_TOKEN || process.env.HF_API_TOKEN;
+const GROQ_API_KEY = process.env.GROQ_API_KEY;
 const nllbLangMap = {
     'en': 'eng_Latn',
     'es': 'spa_Latn',
@@ -77,7 +78,7 @@ function loadSslOptions() {
     // Generate a temporary self-signed certificate for localhost
     console.warn('No SSL cert/key provided. Generating a temporary self-signed certificate for https://localhost:' + SSL_PORT);
     const attrs = [{ name: 'commonName', value: 'localhost' }];
-    const pems = selfsigned.generate(attrs, { days: 365 });
+    const pems = selfsigned.generate(attrs, { days: 365, keySize: 2048 });
     return { key: pems.private, cert: pems.cert };
 }
 
@@ -332,7 +333,25 @@ const translateViaLibreTranslate = async (text, sourceLang, targetLang) => {
     }
 };
 
-// Speech-to-text endpoint using Hugging Face Whisper
+function getTranscriptionFilename(contentType) {
+    const mimeType = (contentType || 'audio/webm').split(';')[0].trim().toLowerCase();
+
+    if (mimeType.includes('mp4') || mimeType.includes('m4a') || mimeType.includes('aac')) {
+        return 'audio.mp4';
+    }
+
+    if (mimeType.includes('webm')) {
+        return 'audio.webm';
+    }
+
+    if (mimeType.includes('wav')) {
+        return 'audio.wav';
+    }
+
+    return 'audio.webm';
+}
+
+// Speech-to-text endpoint using Groq Whisper
 app.post('/api/transcribe', async (req, res) => {
     try {
         const audioBuffer = req.body;
@@ -340,29 +359,40 @@ app.post('/api/transcribe', async (req, res) => {
             return res.status(400).json({ error: 'No audio provided' });
         }
 
-        if (!HUGGINGFACE_API_TOKEN) {
+        if (!GROQ_API_KEY) {
-            return res.status(500).json({ error: 'Hugging Face token not configured' });
+            return res.status(503).json({ error: 'Speech-to-text is temporarily unavailable: GROQ API key is not configured.' });
         }
 
-        const contentType = req.headers['content-type'] || 'application/octet-stream';
+        const contentType = req.headers['content-type'] || 'audio/webm';
-        const response = await axios.post(
+        const mimeType = contentType.split(';')[0].trim().toLowerCase();
-            'https://api-inference.huggingface.co/models/openai/whisper-large-v3',
+        const filename = getTranscriptionFilename(contentType);
-            audioBuffer,
+        const formData = new FormData();
-            {
+
+        formData.append('file', new Blob([audioBuffer], { type: mimeType }), filename);
+        formData.append('model', 'whisper-large-v3-turbo');
+        formData.append('response_format', 'json');
+
+        const response = await fetch('https://api.groq.com/openai/v1/audio/transcriptions', {
+            method: 'POST',
             headers: {
-                    Authorization: `Bearer ${HUGGINGFACE_API_TOKEN}`,
+                Authorization: `Bearer ${GROQ_API_KEY}`
-                    'Content-Type': contentType
             },
-                timeout: 60000
+            body: formData
-            }
+        });
-        );
 
-        const data = response.data;
+        const data = await response.json().catch(() => null);
-        const transcript = Array.isArray(data) ? data[0]?.text : data?.text || '';
+
-        res.json({ transcript });
+        if (!response.ok) {
+            const errorMessage = data?.error?.message || data?.error || 'Speech-to-text failed';
+            console.error('Groq speech-to-text error:', errorMessage);
+            return res.status(response.status).json({ error: errorMessage });
+        }
+
+        const transcript = data?.text || '';
+        return res.json({ transcript });
     } catch (error) {
         console.error('Speech-to-text error:', error.message || error);
-        res.status(500).json({ error: 'Speech-to-text failed' });
+        return res.status(500).json({ error: 'Speech-to-text failed' });
     }
 });