hyperclaw 4.0.0 → 4.0.1

package/dist/server-RBqwE_GN.js

@@ -0,0 +1,4 @@
+const require_chunk = require('./chunk-jS-bbMI5.js');
+const require_server = require('./server-CCI1hv45.js');
+
+exports.startGateway = require_server.startGateway;

package/dist/voice-transcription-CbQBToY0.js

@@ -0,0 +1,138 @@
+const require_chunk = require('./chunk-jS-bbMI5.js');
+const fs_extra = require_chunk.__toESM(require("fs-extra"));
+const path = require_chunk.__toESM(require("path"));
+const os = require_chunk.__toESM(require("os"));
+const https = require_chunk.__toESM(require("https"));
+
+//#region src/services/voice-transcription.ts
+const HC_DIR = path.default.join(os.default.homedir(), ".hyperclaw");
+async function getConfig() {
+	try {
+		const cfg = await fs_extra.default.readJson(path.default.join(HC_DIR, "hyperclaw.json"));
+		const providerId = cfg?.provider?.providerId;
+		const apiKey = cfg?.provider?.apiKey;
+		return {
+			providerId,
+			apiKey
+		};
+	} catch {
+		return {};
+	}
+}
+async function transcribeWithWhisper(buffer, apiKey) {
+	return new Promise((resolve, reject) => {
+		const boundary = "----HyperClaw" + Date.now();
+		const header = [
+			`--${boundary}`,
+			"Content-Disposition: form-data; name=\"file\"; filename=\"audio.ogg\"",
+			"Content-Type: application/octet-stream",
+			"",
+			""
+		].join("\r\n");
+		const footer = `\r\n--${boundary}\r\nContent-Disposition: form-data; name="model"\r\n\r\nwhisper-1\r\n--${boundary}--\r\n`;
+		const body = Buffer.concat([
+			Buffer.from(header, "utf8"),
+			buffer,
+			Buffer.from(footer, "utf8")
+		]);
+		const req = https.default.request({
+			hostname: "api.openai.com",
+			port: 443,
+			path: "/v1/audio/transcriptions",
+			method: "POST",
+			headers: {
+				"Authorization": `Bearer ${apiKey}`,
+				"Content-Type": `multipart/form-data; boundary=${boundary}`,
+				"Content-Length": body.length
+			}
+		}, (res) => {
+			let data = "";
+			res.on("data", (c) => data += c);
+			res.on("end", () => {
+				try {
+					const j = JSON.parse(data);
+					resolve(j.text?.trim() || "[No transcription]");
+				} catch {
+					resolve(`[Transcription error: ${data.slice(0, 100)}]`);
+				}
+			});
+		});
+		req.on("error", reject);
+		req.write(body);
+		req.end();
+	});
+}
+async function transcribeWithGemini(buffer, apiKey) {
+	const base64 = buffer.toString("base64");
+	const payload = JSON.stringify({
+		contents: [{ parts: [{ text: "Transcribe this audio to text. Output only the transcription, no other text." }, { inlineData: {
+			mimeType: "audio/ogg",
+			data: base64
+		} }] }],
+		generationConfig: { maxOutputTokens: 1024 }
+	});
+	return new Promise((resolve, reject) => {
+		const req = https.default.request({
+			hostname: "generativelanguage.googleapis.com",
+			port: 443,
+			path: "/v1beta/models/gemini-2.0-flash:generateContent?key=" + encodeURIComponent(apiKey),
+			method: "POST",
+			headers: {
+				"Content-Type": "application/json",
+				"Content-Length": Buffer.byteLength(payload)
+			}
+		}, (res) => {
+			let data = "";
+			res.on("data", (c) => data += c);
+			res.on("end", () => {
+				try {
+					const j = JSON.parse(data);
+					const text = j.candidates?.[0]?.content?.parts?.[0]?.text?.trim();
+					resolve(text || "[No transcription]");
+				} catch {
+					resolve(`[Transcription error: ${data.slice(0, 100)}]`);
+				}
+			});
+		});
+		req.on("error", reject);
+		req.write(payload);
+		req.end();
+	});
+}
+/**
+
+* Transcribe audio using configured provider or fallbacks.
+
+* Providers: OpenAI (Whisper), Google (Gemini), OpenRouter.
+
+* Env: OPENAI_API_KEY, WHISPER_API_KEY, GOOGLE_AI_API_KEY.
+
+*/
+async function transcribeVoiceNote(audioPathOrBuffer, apiKey) {
+	let buffer;
+	if (typeof audioPathOrBuffer === "string") buffer = await fs_extra.default.readFile(audioPathOrBuffer);
+	else buffer = audioPathOrBuffer;
+	const cfg = await getConfig();
+	const openaiKey = apiKey || process.env.OPENAI_API_KEY || process.env.WHISPER_API_KEY || (cfg.providerId === "openai" || cfg.providerId === "openrouter" ? cfg.apiKey : "");
+	const googleKey = process.env.GOOGLE_AI_API_KEY || (cfg.providerId === "google" ? cfg.apiKey : "");
+	if (cfg.providerId === "google" && googleKey) try {
+		return await transcribeWithGemini(buffer, googleKey);
+	} catch {}
+	if ((cfg.providerId === "openrouter" || cfg.providerId === "openai") && openaiKey) try {
+		return await transcribeWithWhisper(buffer, openaiKey);
+	} catch {}
+	if (openaiKey) try {
+		return await transcribeWithWhisper(buffer, openaiKey);
+	} catch (e) {
+		return `[Transcription failed: ${e.message}]`;
+	}
+	if (googleKey) try {
+		return await transcribeWithGemini(buffer, googleKey);
+	} catch (e) {
+		return `[Transcription failed: ${e.message}]`;
+	}
+	return "[Voice note — add OPENAI_API_KEY or GOOGLE_AI_API_KEY (or select OpenAI/Google provider in the wizard) for transcription]";
+}
+
+//#endregion
+exports.transcribeVoiceNote = transcribeVoiceNote;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "hyperclaw",
-  "version": "4.0.0",
+  "version": "4.0.1",
   "description": "⚡ HyperClaw — AI Gateway Platform. The Lobster Evolution 🦅",
   "main": "dist/run-main.js",
   "bin": {