hyperclaw 5.2.6 → 5.2.8

package/dist/vision-tools-DlnxaZjg.js

@@ -0,0 +1,51 @@
+const require_chunk = require('./chunk-jS-bbMI5.js');
+
+//#region packages/core/src/agent/vision-tools.ts
+function getVisionTools(opts = {}) {
+	const { apiKey = "", provider = "anthropic" } = opts;
+	return [{
+		name: "analyze_image",
+		description: "Analyze an image using a vision model. Describe scenes, receipts, documents, photos. Supports file path, URL, or data URI.",
+		input_schema: {
+			type: "object",
+			properties: {
+				image: {
+					type: "string",
+					description: "Image path (~/path), URL (https://...), or data:image/...;base64,..."
+				},
+				prompt: {
+					type: "string",
+					description: "What to describe or extract (e.g. \"Describe this scene\", \"List items on this receipt\")"
+				}
+			},
+			required: ["image"]
+		},
+		handler: async (input) => {
+			if (!apiKey) return "Error: No API key configured for vision. Set provider.apiKey or run hyperclaw auth add.";
+			const image = input.image?.trim();
+			const prompt = input.prompt?.trim() || "Describe this image concisely.";
+			if (!image) return "Error: image is required";
+			try {
+				const { analyzeImage } = await Promise.resolve().then(() => require("./vision-DV2ubVjr.js"));
+				return await analyzeImage(image, prompt, apiKey, provider);
+			} catch (e) {
+				return `Error: ${e.message}`;
+			}
+		}
+	}];
+}
+var init_vision_tools = require_chunk.__esm({ "packages/core/src/agent/vision-tools.ts"() {} });
+
+//#endregion
+Object.defineProperty(exports, 'getVisionTools', {
+  enumerable: true,
+  get: function () {
+    return getVisionTools;
+  }
+});
+Object.defineProperty(exports, 'init_vision_tools', {
+  enumerable: true,
+  get: function () {
+    return init_vision_tools;
+  }
+});

package/dist/voice-transcription-BolTT_HG.js

@@ -0,0 +1,170 @@
+const require_chunk = require('./chunk-jS-bbMI5.js');
+const require_paths = require('./paths-AIyBxIzm.js');
+const require_paths$1 = require('./paths-DPovhojT.js');
+const fs_extra = require_chunk.__toESM(require("fs-extra"));
+const path = require_chunk.__toESM(require("path"));
+const https = require_chunk.__toESM(require("https"));
+
+//#region src/services/voice-transcription.ts
+require_paths$1.init_paths();
+const MAX_AUDIO_BYTES = 25 * 1024 * 1024;
+const ALLOWED_AUDIO_EXTENSIONS = new Set([
+	".ogg",
+	".oga",
+	".mp3",
+	".wav",
+	".m4a",
+	".mp4",
+	".mpeg",
+	".webm"
+]);
+function sanitizeForError(value) {
+	return String(value ?? "").replace(/[\r\n\t]+/g, " ").slice(0, 160);
+}
+async function loadAudioInput(audioPathOrBuffer) {
+	if (Buffer.isBuffer(audioPathOrBuffer)) {
+		if (audioPathOrBuffer.length === 0) throw new Error("Audio input is empty");
+		if (audioPathOrBuffer.length > MAX_AUDIO_BYTES) throw new Error(`Audio input exceeds ${MAX_AUDIO_BYTES} bytes`);
+		return audioPathOrBuffer;
+	}
+	const resolvedPath = path.default.resolve(audioPathOrBuffer);
+	const ext = path.default.extname(resolvedPath).toLowerCase();
+	if (!ALLOWED_AUDIO_EXTENSIONS.has(ext)) throw new Error(`Unsupported audio file type: ${ext || "unknown"}`);
+	const buf = await fs_extra.default.readFile(resolvedPath);
+	if (buf.length === 0) throw new Error("Audio input is empty");
+	if (buf.length > MAX_AUDIO_BYTES) throw new Error(`Audio input exceeds ${MAX_AUDIO_BYTES} bytes`);
+	return buf;
+}
+async function getConfig() {
+	try {
+		const cfg = await fs_extra.default.readJson(require_paths.getConfigPath());
+		const providerId = cfg?.provider?.providerId;
+		const apiKey = cfg?.provider?.apiKey;
+		return {
+			providerId,
+			apiKey
+		};
+	} catch {
+		return {};
+	}
+}
+async function transcribeWithWhisper(buffer, apiKey) {
+	return new Promise((resolve, reject) => {
+		const boundary = "----HyperClaw" + Date.now();
+		const header = [
+			`--${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;
+	try {
+		buffer = await loadAudioInput(audioPathOrBuffer);
+	} catch (e) {
+		return `[Transcription failed: ${sanitizeForError(e?.message)}]`;
+	}
+	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: ${sanitizeForError(e?.message)}]`;
+	}
+	if (googleKey) try {
+		return await transcribeWithGemini(buffer, googleKey);
+	} catch (e) {
+		return `[Transcription failed: ${sanitizeForError(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/dist/voice-transcription-j1jpOpH2.js

@@ -0,0 +1,170 @@
+const require_chunk = require('./chunk-jS-bbMI5.js');
+const require_paths = require('./paths-AIyBxIzm.js');
+const require_paths$1 = require('./paths-DPovhojT.js');
+const fs_extra = require_chunk.__toESM(require("fs-extra"));
+const path = require_chunk.__toESM(require("path"));
+const https = require_chunk.__toESM(require("https"));
+
+//#region src/services/voice-transcription.ts
+require_paths$1.init_paths();
+const MAX_AUDIO_BYTES = 25 * 1024 * 1024;
+const ALLOWED_AUDIO_EXTENSIONS = new Set([
+	".ogg",
+	".oga",
+	".mp3",
+	".wav",
+	".m4a",
+	".mp4",
+	".mpeg",
+	".webm"
+]);
+function sanitizeForError(value) {
+	return String(value ?? "").replace(/[\r\n\t]+/g, " ").slice(0, 160);
+}
+async function loadAudioInput(audioPathOrBuffer) {
+	if (Buffer.isBuffer(audioPathOrBuffer)) {
+		if (audioPathOrBuffer.length === 0) throw new Error("Audio input is empty");
+		if (audioPathOrBuffer.length > MAX_AUDIO_BYTES) throw new Error(`Audio input exceeds ${MAX_AUDIO_BYTES} bytes`);
+		return audioPathOrBuffer;
+	}
+	const resolvedPath = path.default.resolve(audioPathOrBuffer);
+	const ext = path.default.extname(resolvedPath).toLowerCase();
+	if (!ALLOWED_AUDIO_EXTENSIONS.has(ext)) throw new Error(`Unsupported audio file type: ${ext || "unknown"}`);
+	const buf = await fs_extra.default.readFile(resolvedPath);
+	if (buf.length === 0) throw new Error("Audio input is empty");
+	if (buf.length > MAX_AUDIO_BYTES) throw new Error(`Audio input exceeds ${MAX_AUDIO_BYTES} bytes`);
+	return buf;
+}
+async function getConfig() {
+	try {
+		const cfg = await fs_extra.default.readJson(require_paths.getConfigPath());
+		const providerId = cfg?.provider?.providerId;
+		const apiKey = cfg?.provider?.apiKey;
+		return {
+			providerId,
+			apiKey
+		};
+	} catch {
+		return {};
+	}
+}
+async function transcribeWithWhisper(buffer, apiKey) {
+	return new Promise((resolve, reject) => {
+		const boundary = "----HyperClaw" + Date.now();
+		const header = [
+			`--${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;
+	try {
+		buffer = await loadAudioInput(audioPathOrBuffer);
+	} catch (e) {
+		return `[Transcription failed: ${sanitizeForError(e?.message)}]`;
+	}
+	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: ${sanitizeForError(e?.message)}]`;
+	}
+	if (googleKey) try {
+		return await transcribeWithGemini(buffer, googleKey);
+	} catch (e) {
+		return `[Transcription failed: ${sanitizeForError(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": "5.2.6",
+  "version": "5.2.8",
   "description": "⚡ HyperClaw — AI Gateway Platform. The Lobster Evolution 🦅",
   "main": "dist/run-main.js",
   "bin": {
@@ -100,6 +100,8 @@
     "fs-extra": "^11.2.0",
     "gradient-string": "^2.0.2",
     "inquirer": "^8.2.7",
+    "marked": "^15.0.12",
+    "marked-terminal": "^7.3.0",
     "node-cron": "^3.0.3",
     "ora": "^5.4.1",
     "tar": "^7.0.0",
@@ -119,6 +121,7 @@
     "@types/figlet": "^1.5.8",
     "@types/fs-extra": "^11.0.4",
     "@types/inquirer": "^8.2.10",
+    "@types/marked-terminal": "^6.1.1",
     "@types/node": "^22.0.0",
     "@types/ws": "^8.5.8",
     "@vitest/coverage-v8": "^1.6.0",

package/static/chat.html

@@ -139,10 +139,10 @@
 <!-- INPUT -->
 <div class="input-wrap">
   <form class="input-row" id="form">
-    <textarea id="input" rows="1" placeholder="Message HyperClaw… (Enter to send, Shift+Enter for new line)" autocomplete="off"></textarea>
+    <textarea id="input" rows="1" placeholder="Your AI assistant awaits — type a message and press Enter to send · HyperClaw" autocomplete="off"></textarea>
     <button type="submit" id="send-btn">↑</button>
   </form>
-  <div class="input-hint">🦅 Hyper · <span id="model-hint">—</span></div>
+  <div class="input-hint">🦅 HyperClaw · <span id="model-hint">—</span></div>
 </div>
 
 <script>
@@ -270,17 +270,23 @@
     if (last) last.appendChild(t);
   }
 
+  let workingTicker = null;
   function addTyping() {
     welcomeEl.style.display = 'none';
     const row = document.createElement('div');
     row.className = 'msg-row assistant';
     row.id = 'typing-row';
-    row.innerHTML = '<div class="msg-avatar">🦅</div><div class="msg-bubble" style="padding:14px 18px"><span class="typing-dot"></span><span class="typing-dot"></span><span class="typing-dot"></span></div>';
+    let sec = 0;
+    row.innerHTML = '<div class="msg-avatar">🦅</div><div class="msg-bubble" style="padding:14px 18px"><span class="typing-dot"></span><span class="typing-dot"></span><span class="typing-dot"></span> <span id="working-text" style="margin-left:8px;color:var(--text2);font-size:12px">Working (0s)</span></div>';
     messagesEl.appendChild(row);
     messagesEl.scrollTop = messagesEl.scrollHeight;
+    workingTicker = setInterval(() => {
+      sec++; const el = document.getElementById('working-text'); if (el) el.textContent = 'Working (' + sec + 's)';
+    }, 1000);
   }
 
   function removeTyping() {
+    if (workingTicker) { clearInterval(workingTicker); workingTicker = null; }
     document.getElementById('typing-row')?.remove();
   }