@andresaya/edge-tts 1.7.2 → 1.8.1

package/dist/services/EdgeTTS.js

@@ -30,30 +30,88 @@ class EdgeTTS {
     output_format = 'audio-24khz-48kbitrate-mono-mp3';
     word_boundaries = [];
     ws;
+    async normalizeVoices(data) {
+        const out = [];
+        for (const v of data || []) {
+            const short = v?.ShortName || "";
+            const locale = v?.Locale || "";
+            // base: remove locale prefix and Neural/NeuralHD suffix
+            let base = short.replace(/^[a-z]{2}-[A-Z]{2}-/, "");
+            base = base.replace(/NeuralHD$/, "").replace(/Neural$/, "").trim();
+            // VoiceType: if NeuralHD/Neural in Name or ShortName
+            const mix = `${v?.Name || ""} ${short}`;
+            const voiceType = v?.VoiceType || (/NeuralHD/i.test(mix) ? "NeuralHD" : "Neural");
+            // LocaleName: prefer LocaleName -> LanguageName -> locale
+            const localeName = v?.LocaleName || (locale || null);
+            // DisplayName: prefer DisplayName -> FriendlyName -> base -> short
+            let display = v?.DisplayName || v?.FriendlyName || base || short;
+            display = display.replace(/^Microsoft\s+/i, "");
+            display = display.split(" - ")[0].trim();
+            display = display.replace(/\s*Online\s*\(Natural\)\s*/i, " ");
+            display = display.replace(/\s*Online\s*/i, " ");
+            display = display.replace(/\s+/g, " ").trim();
+            // VoiceTag parsing
+            const tag = (v?.VoiceTag && typeof v.VoiceTag === "object") ? v.VoiceTag : {};
+            const tailored = Array.isArray(tag.TailoredScenarios)
+                ? tag.TailoredScenarios
+                : (Array.isArray(tag.ContentCategories) ? tag.ContentCategories : []);
+            const personalities = Array.isArray(tag.VoicePersonalities)
+                ? tag.VoicePersonalities
+                : [];
+            out.push({
+                Name: short || (v?.Name || ""),
+                DisplayName: display,
+                LocalName: display,
+                ShortName: short || (v?.Name || ""),
+                Gender: v?.Gender ?? null,
+                Locale: locale || null,
+                LocaleName: localeName,
+                SecondaryLocaleList: Array.isArray(v?.SecondaryLocaleList) ? v.SecondaryLocaleList : [],
+                VoiceType: voiceType,
+                VoiceTag: {
+                    TailoredScenarios: tailored,
+                    VoicePersonalities: personalities,
+                },
+                FriendlyName: `${display} (${voiceType}) - ${localeName}`,
+            });
+        }
+        return out;
+    }
     async getVoices() {
         const secMsGEC = await this.generateSecMsGec(constants_1.Constants.TRUSTED_CLIENT_TOKEN);
-        const httpsAgent = new https_1.default.Agent({
-            rejectUnauthorized: false,
-        });
-        const response = await new Promise((resolve, reject) => {
-            const req = https_1.default.request(`${constants_1.Constants.VOICES_URL}?Ocp-Apim-Subscription-Key=${constants_1.Constants.TRUSTED_CLIENT_TOKEN}&Sec-MS-GEC=${secMsGEC}&Sec-MS-GEC-Version=${constants_1.Constants.VERSION_MS_GEC}`, {
-                headers: constants_1.Constants.getBaseHeaders(),
-                agent: httpsAgent
-            }, (res) => {
-                let data = '';
-                res.on('data', (chunk) => (data += chunk));
-                res.on('end', () => resolve(new Response(data, { status: res.statusCode || 200 })));
+        const httpsAgent = new https_1.default.Agent({ rejectUnauthorized: false });
+        const url = `${constants_1.Constants.VOICES_URL}` +
+            `?TrustedClientToken=${constants_1.Constants.TRUSTED_CLIENT_TOKEN}` +
+            `&Sec-MS-GEC=${secMsGEC}` +
+            `&Sec-MS-GEC-Version=${constants_1.Constants.VERSION_MS_GEC}`;
+        const headers = {
+            ...constants_1.Constants.getBaseHeaders(),
+            "Accept-Encoding": "identity", // evita gzip/br/zstd
+        };
+        const voicesRaw = await new Promise((resolve, reject) => {
+            const req = https_1.default.request(url, { method: "GET", headers, agent: httpsAgent }, (res) => {
+                const chunks = [];
+                res.on("data", (chunk) => chunks.push(buffer_1.Buffer.isBuffer(chunk) ? chunk : buffer_1.Buffer.from(chunk)));
+                res.on("error", reject);
+                res.on("end", () => {
+                    const body = buffer_1.Buffer.concat(chunks).toString("utf8");
+                    try {
+                        const parsed = JSON.parse(body);
+                        // soporta array directo o { voices: [...] }
+                        const voices = Array.isArray(parsed)
+                            ? parsed
+                            : (parsed.voices || parsed.Voices || []);
+                        resolve(Array.isArray(voices) ? voices : []);
+                    }
+                    catch (e) {
+                        reject(new Error("JSON inválido: " + (e?.message || String(e))));
+                    }
+                });
             });
-            req.on('error', reject);
+            req.on("error", reject);
             req.end();
         });
-        const data = await response.json();
-        return data.map((voice) => {
-            voice.FriendlyName = voice.FriendlyName || voice.LocalName;
-            delete voice.SampleRateHertz;
-            delete voice.Status;
-            return voice;
-        });
+        return this.normalizeVoices(voicesRaw);
     }
     async getVoicesByLanguage(locale) {
         const voices = await this.getVoices();
@@ -114,7 +172,7 @@ class EdgeTTS {
         return new Promise((resolve, reject) => {
             this.audio_stream = [];
             const reqId = this.generateUUID();
-            const url = `${constants_1.Constants.WSS_URL}?Ocp-Apim-Subscription-Key=${constants_1.Constants.TRUSTED_CLIENT_TOKEN}&Sec-MS-GEC=${secMsGEC}&Sec-MS-GEC-Version=${constants_1.Constants.VERSION_MS_GEC}&ConnectionId=${reqId}`;
+            const url = `${constants_1.Constants.WSS_URL}?TrustedClientToken=${constants_1.Constants.TRUSTED_CLIENT_TOKEN}&Sec-MS-GEC=${secMsGEC}&Sec-MS-GEC-Version=${constants_1.Constants.VERSION_MS_GEC}&ConnectionId=${reqId}`;
             this.ws = new ws_1.default(url, {
                 headers: constants_1.Constants.getBaseHeaders(),
                 rejectUnauthorized: false
@@ -161,31 +219,6 @@ class EdgeTTS {
             });
         });
     }
-    detectSSML(content) {
-        const trimmedContent = content.trim();
-        const looksLikeSSML = /^<\?xml|^<speak/i.test(trimmedContent);
-        if (!looksLikeSSML) {
-            return { isValid: true, isSSML: false };
-        }
-        const errors = [];
-        const hasSpeakTag = /<speak\b[^>]*>[\s\S]*<\/speak>/i.test(trimmedContent);
-        const hasVoiceTag = /<voice\b[^>]*>[\s\S]*<\/voice>/i.test(trimmedContent);
-        if (!hasSpeakTag) {
-            throw new Error('Invalid SSML: Missing <speak> tag');
-        }
-        if (!hasVoiceTag) {
-            throw new Error('Invalid SSML: Missing <voice> tag');
-        }
-        const hasCorrectNamespace = /xmlns="http:\/\/www\.w3\.org\/2001\/10\/synthesis"/i.test(trimmedContent);
-        if (!hasCorrectNamespace && hasSpeakTag) {
-            throw new Error('Invalid SSML: Missing or incorrect namespace declaration');
-        }
-        return {
-            isValid: errors.length === 0,
-            isSSML: hasSpeakTag || hasVoiceTag,
-            errors: errors.length > 0 ? errors : undefined
-        };
-    }
     escapeXML(text) {
         return text
             .replace(/&/g, '&amp;')
@@ -195,48 +228,17 @@ class EdgeTTS {
             .replace(/'/g, '&apos;');
     }
     getSSML(content, voice, options = {}) {
-        const inputType = options.inputType || 'auto';
-        let treatAsSSML = false;
-        if (inputType === 'ssml') {
-            treatAsSSML = true;
-        }
-        else if (inputType === 'text') {
-            treatAsSSML = false;
-        }
-        else { // 'auto'
-            const detection = this.detectSSML(content);
-            treatAsSSML = detection.isSSML;
-            if (detection.isSSML) {
-                console.log('→ Detected SSML input');
-                if (!detection.isValid) {
-                    console.warn('⚠ SSML validation warnings:', detection.errors);
-                }
-            }
-            else {
-                console.log('→ Detected plain text input');
-            }
-        }
-        if (treatAsSSML) {
-            let ssml = content.trim();
-            if (!ssml.includes('xmlns=')) {
-                ssml = ssml.replace(/<speak([^>]*)>/i, '<speak$1 xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="https://www.w3.org/2001/mstts">');
-            }
-            if (!/<voice\b[^>]*>/i.test(ssml) && voice) {
-                ssml = ssml.replace(/(<speak[^>]*>)([\s\S]*?)(<\/speak>)/i, `$1<voice name="${voice}">$2</voice>$3`);
-            }
-            return ssml;
-        }
         const pitch = this.validatePitch(options.pitch ?? 0);
         const rate = this.validateRate(options.rate ?? 0);
         const volume = this.validateVolume(options.volume ?? 0);
         const escapedText = this.escapeXML(content);
-        return `<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="https://www.w3.org/2001/mstts" xml:lang="en-US">
-                    <voice name="${voice}">
-                        <prosody pitch="${pitch}" rate="${rate}" volume="${volume}">
-                            ${escapedText}
-                        </prosody>
-                    </voice>
-                </speak>
+        return `<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="https://www.w3.org/2001/mstts" xml:lang="en-US">
+                    <voice name="${voice}">
+                        <prosody pitch="${pitch}" rate="${rate}" volume="${volume}">
+                            ${escapedText}
+                        </prosody>
+                    </voice>
+                </speak>
         `;
     }
     nowRFC1123(timeZone = 'UTC') {
@@ -266,7 +268,7 @@ class EdgeTTS {
         this.audio_stream = [];
         const reqId = this.generateUUID();
         const secMsGEC = await this.generateSecMsGec(constants_1.Constants.TRUSTED_CLIENT_TOKEN);
-        const url = `${constants_1.Constants.WSS_URL}?Ocp-Apim-Subscription-Key=${constants_1.Constants.TRUSTED_CLIENT_TOKEN}&Sec-MS-GEC=${secMsGEC}&Sec-MS-GEC-Version=${constants_1.Constants.VERSION_MS_GEC}&ConnectionId=${reqId}`;
+        const url = `${constants_1.Constants.WSS_URL}?TrustedClientToken=${constants_1.Constants.TRUSTED_CLIENT_TOKEN}&Sec-MS-GEC=${secMsGEC}&Sec-MS-GEC-Version=${constants_1.Constants.VERSION_MS_GEC}&ConnectionId=${reqId}`;
         this.ws = new ws_1.default(url, {
             headers: constants_1.Constants.getBaseHeaders(),
             rejectUnauthorized: false
@@ -484,7 +486,16 @@ class EdgeTTS {
         if (this.audio_stream.length === 0) {
             throw new Error("No audio data available. Did you run synthesize() first?");
         }
-        return buffer_1.Buffer.concat(this.audio_stream);
+        // When the service cannot speak the text it answers with a single empty
+        // chunk, so the stream is not empty but the audio is. Checking the byte
+        // count rather than the chunk count is what catches that.
+        const buffer = buffer_1.Buffer.concat(this.audio_stream);
+        if (buffer.length === 0) {
+            throw new Error("The service returned no audio. Voices are locale specific, so a voice cannot " +
+                "read text written in another script. Check that the voice matches the language " +
+                "of the text.");
+        }
+        return buffer;
     }
     async saveMetadata(outputPath) {
         if (this.word_boundaries.length === 0) {

package/package.json

@@ -1,54 +1,54 @@
-{
-  "name": "@andresaya/edge-tts",
-  "version": "1.7.2",
-  "description": "Edge TTS is a package that allows access to the online text-to-speech service used by Microsoft Edge without the need for Microsoft Edge, Windows, or an API key.",
-  "main": "dist/index.js",
-  "browser": "dist/browser/edge-tts.esm.js",
-  "unpkg": "dist/browser/edge-tts.umd.min.js",
-  "jsdelivr": "dist/browser/edge-tts.umd.min.js",
-  "keywords": [
-    "edge",
-    "tts",
-    "text-to-speech",
-    "microsoft",
-    "microsoft edge",
-    "speech-synthesis"
-  ],
-  "author": "Andres Aya",
-  "license": "GPL-3.0-only",
-  "homepage": "https://github.com/andresayac/edge-tts",
-  "bugs": {
-    "url": "https://github.com/andresayac/edge-tts/issues"
-  },
-  "scripts": {
-    "build": "tsc",
-    "build:bin": "bun scripts/build.ts",
-    "build:browser": "bun scripts/build-browser.ts",
-    "prepublishOnly": "npm run build && npm run build:bin && npm run build:browser"
-  },
-  "devDependencies": {
-    "@types/bun": "latest",
-    "@types/ws": "^8.18.1",
-    "esbuild": "^0.25.12",
-    "typescript": "^5.6.2"
-  },
-  "dependencies": {
-    "@types/uuid": "^10.0.0",
-    "commander": "^12.1.0",
-    "events": "^3.3.0",
-    "fs": "^0.0.1-security",
-    "ws": "^8.18.3"
-  },
-  "peerDependencies": {
-    "typescript": "^5.0.0"
-  },
-  "files": [
-    "dist"
-  ],
-  "bin": {
-    "edge-tts": "./dist/cli/edge-tts.js"
-  },
-  "engines": {
-    "bun": ">=0.5.0"
-  }
-}
+{
+  "name": "@andresaya/edge-tts",
+  "version": "1.8.1",
+  "description": "Edge TTS is a package that allows access to the online text-to-speech service used by Microsoft Edge without the need for Microsoft Edge, Windows, or an API key.",
+  "main": "dist/index.js",
+  "browser": "dist/browser/edge-tts.esm.js",
+  "unpkg": "dist/browser/edge-tts.umd.min.js",
+  "jsdelivr": "dist/browser/edge-tts.umd.min.js",
+  "keywords": [
+    "edge",
+    "tts",
+    "text-to-speech",
+    "microsoft",
+    "microsoft edge",
+    "speech-synthesis"
+  ],
+  "author": "Andres Aya",
+  "license": "GPL-3.0-only",
+  "homepage": "https://github.com/andresayac/edge-tts",
+  "bugs": {
+    "url": "https://github.com/andresayac/edge-tts/issues"
+  },
+  "scripts": {
+    "build": "tsc",
+    "build:bin": "bun scripts/build.ts",
+    "build:browser": "bun scripts/build-browser.ts",
+    "prepublishOnly": "npm run build && npm run build:bin && npm run build:browser"
+  },
+  "devDependencies": {
+    "@types/bun": "latest",
+    "@types/ws": "^8.18.1",
+    "esbuild": "^0.25.12",
+    "typescript": "^5.6.2"
+  },
+  "dependencies": {
+    "@types/uuid": "^10.0.0",
+    "commander": "^12.1.0",
+    "events": "^3.3.0",
+    "fs": "^0.0.1-security",
+    "ws": "^8.18.3"
+  },
+  "peerDependencies": {
+    "typescript": "^5.0.0"
+  },
+  "files": [
+    "dist"
+  ],
+  "bin": {
+    "edge-tts": "dist/cli/edge-tts.js"
+  },
+  "engines": {
+    "bun": ">=0.5.0"
+  }
+}