mulmocast 2.1.8 → 2.1.10

package/lib/actions/bundle.js

@@ -6,6 +6,7 @@ import { mkdir } from "../utils/file.js";
 import { ZipBuilder } from "../utils/zip.js";
 import { bundleTargetLang } from "../utils/const.js";
 import { createSilentAudio } from "../utils/ffmpeg_utils.js";
+import { silentMp3 } from "../utils/context.js";
 const downloadFile = async (url, destPath) => {
     const response = await fetch(url);
     if (!response.ok) {
@@ -53,7 +54,7 @@ const imageSourceMappings = [
 ];
 export const mulmoViewerBundle = async (context) => {
     const isZip = true;
-    const dir = path.resolve(context.fileDirs.fileName);
+    const dir = context.fileDirs.outDirPath;
     mkdir(dir);
     const zipper = new ZipBuilder(path.resolve(dir, zipFileName));
     // text
@@ -67,18 +68,24 @@ export const mulmoViewerBundle = async (context) => {
     // audio
     for (const lang of bundleTargetLang) {
         const audios = listLocalizedAudioPaths({ ...context, lang });
-        audios.forEach((audio, index) => {
+        await Promise.all(audios.map(async (audio, index) => {
             if (audio) {
                 const fileName = path.basename(audio ?? "");
                 if (resultJson[index] && resultJson[index].audioSources) {
                     resultJson[index].audioSources[lang] = fileName;
                 }
-                if (fs.existsSync(audio)) {
+                if (fileName === "silent300.mp3") {
+                    // Download from GitHub URL
+                    const destPath = path.resolve(dir, fileName);
+                    await downloadFile(silentMp3, destPath);
+                    zipper.addFile(destPath, fileName);
+                }
+                else if (fs.existsSync(audio)) {
                     fs.copyFileSync(audio, path.resolve(dir, fileName));
                     zipper.addFile(audio, fileName);
                 }
             }
-        });
+        }));
     }
     // image, movie
     context.studio.beats.forEach((image, index) => {

package/lib/agents/tts_gemini_agent.js

@@ -3,9 +3,16 @@ import { GoogleGenAI } from "@google/genai";
 import { provider2TTSAgent } from "../utils/provider2agent.js";
 import { agentIncorrectAPIKeyError, apiKeyMissingError, agentGenerationError, audioAction, audioFileTarget, getGenAIErrorReason, } from "../utils/error_cause.js";
 import { pcmToMp3 } from "../utils/ffmpeg_utils.js";
+const getPrompt = (text, instructions) => {
+    // https://ai.google.dev/gemini-api/docs/speech-generation?hl=ja#controllable
+    if (instructions) {
+        return `### DIRECTOR'S NOTES\n${instructions}\n\n#### TRANSCRIPT\n${text}`;
+    }
+    return text;
+};
 export const ttsGeminiAgent = async ({ namedInputs, params, config, }) => {
     const { text } = namedInputs;
-    const { model, voice, suppressError } = params;
+    const { model, voice, suppressError, instructions } = params;
     const apiKey = config?.apiKey;
     if (!apiKey) {
         throw new Error("Google GenAI API key is required (GEMINI_API_KEY)", {
@@ -16,7 +23,7 @@ export const ttsGeminiAgent = async ({ namedInputs, params, config, }) => {
         const ai = new GoogleGenAI({ apiKey });
         const response = await ai.models.generateContent({
             model: model ?? provider2TTSAgent.gemini.defaultModel,
-            contents: [{ parts: [{ text }] }],
+            contents: [{ parts: [{ text: getPrompt(text, instructions) }] }],
             config: {
                 responseModalities: ["AUDIO"],
                 speechConfig: {

package/lib/types/agent.d.ts

@@ -122,6 +122,7 @@ export type KotodamaTTSAgentParams = TTSAgentParams & {
 export type GoogleTTSAgentParams = TTSAgentParams & {
     speed: number;
     model: string;
+    instructions: string;
 };
 export type ElevenlabsTTSAgentParams = TTSAgentParams & {
     model: string;

package/lib/utils/context.d.ts

@@ -1,4 +1,5 @@
 import type { MulmoStudioBeat, MulmoScript, MulmoPresentationStyle, MulmoStudioMultiLingual, FileObject } from "../types/type.js";
+export declare const silentMp3 = "https://github.com/receptron/mulmocast-cli/raw/refs/heads/main/assets/audio/silent300.mp3";
 export declare const createStudioData: (_mulmoScript: MulmoScript, fileName: string, videoCaptionLang?: string, presentationStyle?: MulmoPresentationStyle | null) => {
     script: {
         $mulmocast: {

package/lib/utils/context.js

@@ -3,6 +3,7 @@ import { readMulmoScriptFile, fetchMulmoScriptFile, isFile } from "./file.js";
 import { beatId, multiLingualObjectToArray } from "./utils.js";
 import { mulmoStudioSchema, mulmoCaptionParamsSchema, mulmoPresentationStyleSchema } from "../types/schema.js";
 import { MulmoPresentationStyleMethods, MulmoScriptMethods, MulmoStudioMultiLingualMethod } from "../methods/index.js";
+export const silentMp3 = "https://github.com/receptron/mulmocast-cli/raw/refs/heads/main/assets/audio/silent300.mp3";
 const mulmoCredit = (speaker) => {
     return {
         id: "mulmo_credit",
@@ -19,7 +20,7 @@ const mulmoCredit = (speaker) => {
             type: "audio",
             source: {
                 kind: "url",
-                url: "https://github.com/receptron/mulmocast-cli/raw/refs/heads/main/assets/audio/silent300.mp3",
+                url: silentMp3,
             },
         },
     };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "mulmocast",
-  "version": "2.1.8",
+  "version": "2.1.10",
   "description": "",
   "type": "module",
   "main": "lib/index.node.js",
@@ -74,7 +74,7 @@
   "homepage": "https://github.com/receptron/mulmocast-cli#readme",
   "dependencies": {
     "@google-cloud/text-to-speech": "^6.4.0",
-    "@google/genai": "^1.33.0",
+    "@google/genai": "^1.34.0",
     "@graphai/anthropic_agent": "^2.0.12",
     "@graphai/browserless_agent": "^2.0.1",
     "@graphai/gemini_agent": "^2.0.1",
@@ -86,7 +86,7 @@
     "@graphai/vanilla_node_agents": "^2.0.4",
     "@inquirer/input": "^4.3.0",
     "@inquirer/select": "^4.4.1",
-    "@modelcontextprotocol/sdk": "^1.24.3",
+    "@modelcontextprotocol/sdk": "^1.25.1",
     "@mozilla/readability": "^0.6.0",
     "@tavily/core": "^0.5.11",
     "archiver": "^7.0.1",
@@ -102,7 +102,7 @@
     "replicate": "^1.4.0",
     "yaml": "^2.8.2",
     "yargs": "^18.0.0",
-    "zod": "^4.1.13"
+    "zod": "^4.2.1"
   },
   "devDependencies": {
     "@receptron/test_utils": "^2.0.3",
@@ -117,7 +117,7 @@
     "prettier": "^3.7.4",
     "tsx": "^4.21.0",
     "typescript": "^5.9.3",
-    "typescript-eslint": "^8.49.0"
+    "typescript-eslint": "^8.50.0"
   },
   "engines": {
     "node": ">=20.0.0"

package/scripts/test/test_audio_gemini.json

@@ -14,6 +14,14 @@
         "provider": "gemini",
         "model": "gemini-2.5-pro-preview-tts",
         "voiceId": "Puck"
+      },
+      "Presenter2": {
+        "provider": "gemini",
+        "model": "gemini-2.5-pro-preview-tts",
+        "voiceId": "leda",
+        "speechOptions": {
+          "instruction": "Role: You are a classic Tsundere character. Tone: Sharp, impatient, and defensive, but with underlying hesitation that suggests you actually care deeply. Voice Quality: Higher pitch, energetic, and slightly haughty."
+        }
       }
     }
   },
@@ -52,7 +60,7 @@
       }
     },
     {
-      "speaker": "Presenter",
+      "speaker": "Presenter2",
       "text": "Hello, I'm a presenter. I have a whisper instruction.",
       "speechOptions": {
         "instruction": "Whisper softly, like a pillow talk."

package/scripts/test/test_media.json

@@ -60,7 +60,7 @@
       "speaker": "Presenter",
       "text": "",
       "duration": 0.5,
-      "id": "{1A57B3F5-B6CB-4948-96BB-6F018DCCBBD4}",
+      "id": "1A57B3F5-B6CB-4948-96BB-6F018DCCBBD4",
       "image": {
         "type": "textSlide",
         "slide": {