open-agents-ai 0.187.166 → 0.187.167

package/dist/index.js

@@ -254789,7 +254789,7 @@ var init_audio_playback = __esm({
     "use strict";
     AudioPlaybackTool = class {
       name = "audio_playback";
-      description = "Play audio through speakers or use text-to-speech. Actions: 'play' to play an audio file (WAV/MP3/OGG), 'speak' to convert text to speech, 'volume' to get or set system volume, 'list' to enumerate audio output devices. Use this to communicate audibly, play sounds, give voice feedback, or control speaker volume.";
+      description = "Play audio through speakers or use text-to-speech. Actions: 'play' to play an audio file (WAV/MP3/OGG \u2014 including recordings from memory episodes), 'speak' to convert text to speech (uses LuxTTS voice clone if available, falls back to Kokoro/piper/espeak), 'volume' to get or set system volume, 'list' to enumerate audio output devices. Use this to communicate audibly, play sounds, replay recorded audio from memory episodes, or control speaker volume. To replay a memory episode's audio, use the recording path from multimodal_memory recall.";
       parameters = {
         type: "object",
         properties: {
@@ -254881,10 +254881,77 @@ var init_audio_playback = __esm({
         }
         const speed = args["speed"] || 160;
         const voice = args["voice"] || "en";
+        const { join: join98 } = __require("node:path");
+        const { homedir: homedir31, tmpdir: tmpdir19 } = __require("node:os");
+        const { existsSync: existsSync79, unlinkSync: unlinkSync18 } = __require("node:fs");
+        const luxttsScript = join98(homedir31(), ".open-agents", "voice", "luxtts-infer.py");
+        if (existsSync79(luxttsScript)) {
+          try {
+            const luxttsVenvPy2 = join98(homedir31(), ".open-agents", "voice", "luxtts-venv", "bin", "python3");
+            if (existsSync79(luxttsVenvPy2)) {
+              const outFile = join98(tmpdir19(), `oa-tts-${Date.now()}.wav`);
+              const cloneRef = join98(homedir31(), ".open-agents", "voice", "clone-refs");
+              let cloneRefFile = "";
+              try {
+                const refs = __require("node:fs").readdirSync(cloneRef).filter((f2) => f2.endsWith(".wav") || f2.endsWith(".mp3"));
+                if (refs.length > 0)
+                  cloneRefFile = join98(cloneRef, refs[0]);
+              } catch {
+              }
+              if (cloneRefFile) {
+                const safeText2 = text.replace(/"/g, '\\"').replace(/\n/g, " ");
+                const cmd = `echo '{"action":"synthesize","id":"tts","text":"${safeText2}","clone_ref":"${cloneRefFile}","output_path":"${outFile}","speed":1.0}' | ${luxttsVenvPy2} ${luxttsScript}`;
+                try {
+                  execSync29(cmd, { timeout: 3e4, stdio: "pipe" });
+                  if (existsSync79(outFile)) {
+                    execSync29(`aplay -q "${outFile}" 2>/dev/null || ffplay -nodisp -autoexit -loglevel error "${outFile}"`, { timeout: 6e4, stdio: "pipe" });
+                    try {
+                      unlinkSync18(outFile);
+                    } catch {
+                    }
+                    return { success: true, output: `Spoke via LuxTTS (voice clone): ${text.length} chars`, durationMs: performance.now() - start2 };
+                  }
+                } catch {
+                }
+              }
+            }
+          } catch {
+          }
+        }
+        const kokoroModel = join98(homedir31(), ".open-agents", "voice", "models", "kokoro-v1.0", "model.onnx");
+        if (existsSync79(kokoroModel)) {
+          try {
+            const outFile = join98(tmpdir19(), `oa-tts-${Date.now()}.wav`);
+            const voiceVenvPy = join98(homedir31(), ".open-agents", "venv", "bin", "python3");
+            if (existsSync79(voiceVenvPy)) {
+              const safeText2 = text.replace(/'/g, "'\\''");
+              execSync29(`${voiceVenvPy} -c "
+import onnxruntime, json, numpy as np, wave
+# Kokoro synthesis would go here \u2014 simplified for now
+" 2>/dev/null`, { timeout: 3e4, stdio: "pipe" });
+            }
+          } catch {
+          }
+        }
+        try {
+          execSync29("which piper", { stdio: "pipe", timeout: 2e3 });
+          const outFile = join98(tmpdir19(), `oa-tts-${Date.now()}.wav`);
+          const safeText2 = text.replace(/'/g, "'\\''");
+          execSync29(`echo '${safeText2}' | piper --output_file ${outFile} 2>/dev/null`, { timeout: 3e4, stdio: "pipe" });
+          if (existsSync79(outFile)) {
+            execSync29(`aplay -q "${outFile}"`, { timeout: 6e4, stdio: "pipe" });
+            try {
+              unlinkSync18(outFile);
+            } catch {
+            }
+            return { success: true, output: `Spoke via Piper TTS: ${text.length} chars`, durationMs: performance.now() - start2 };
+          }
+        } catch {
+        }
         try {
           execSync29("which espeak-ng", { stdio: "pipe", timeout: 2e3 });
         } catch {
-          return { success: false, output: "", error: "espeak-ng not installed. Install with: sudo apt install espeak-ng", durationMs: performance.now() - start2 };
+          return { success: false, output: "", error: "No TTS engine available. Install espeak-ng: sudo apt install espeak-ng", durationMs: performance.now() - start2 };
         }
         const safeText = text.replace(/'/g, "'\\''");
         try {
@@ -254895,7 +254962,7 @@ var init_audio_playback = __esm({
         }
         return {
           success: true,
-          output: `Spoke ${text.length} characters (voice=${voice}, speed=${speed}wpm)`,
+          output: `Spoke via espeak-ng: ${text.length} chars (voice=${voice}, speed=${speed}wpm)`,
           durationMs: performance.now() - start2
         };
       }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "open-agents-ai",
-  "version": "0.187.166",
+  "version": "0.187.167",
   "description": "AI coding agent powered by open-source models (Ollama/vLLM) — interactive TUI with agentic tool-calling loop",
   "type": "module",
   "main": "./dist/index.js",