mulmocast 1.2.34 → 1.2.36

package/README.md

@@ -106,13 +106,11 @@ OPENAI_API_KEY=your_openai_api_key
 DEFAULT_OPENAI_IMAGE_MODEL=gpt-image-1 # for the advanced image generation model
 ```
 
-#### (Optional) For Google's image generation model
+#### (Optional) For Google's image and TTS.
 ```bash
-GOOGLE_PROJECT_ID=your_google_project_id
+GEMINI_API_KEY=your_google_gemini_api_key
 ```
 
-See also [pre-requisites for Google's image generation model](./docs/pre-requisites-google.md)
-
 #### (Optional) For AI providers
 ```bash
 # For Anthropic Claude (htmlPrompt feature)

package/lib/actions/audio.js

@@ -9,7 +9,6 @@ import { getAudioArtifactFilePath, getAudioFilePath, getOutputStudioFilePath, re
 import { localizedText, settings2GraphAIConfig } from "../utils/utils.js";
 import { text2hash } from "../utils/utils_node.js";
 import { provider2TTSAgent } from "../utils/provider2agent.js";
-import { MulmoPresentationStyleMethods } from "../methods/index.js";
 import { MulmoStudioContextMethods } from "../methods/mulmo_studio_context.js";
 import { MulmoMediaSourceMethods } from "../methods/mulmo_media_source.js";
 dotenv.config({ quiet: true });
@@ -27,15 +26,9 @@ const getAudioPath = (context, beat, audioFile) => {
     }
     return audioFile;
 };
-const getAudioParam = (context, beat, lang) => {
-    const speaker = MulmoPresentationStyleMethods.getSpeaker(context, beat, lang);
-    const speechOptions = { ...speaker.speechOptions, ...beat.speechOptions };
-    const provider = text2SpeechProviderSchema.parse(speaker.provider);
-    return { voiceId: speaker.voiceId, provider, speechOptions, model: speaker.model };
-};
 export const getBeatAudioPath = (text, context, beat, lang) => {
     const audioDirPath = MulmoStudioContextMethods.getAudioDirPath(context);
-    const { voiceId, provider, speechOptions, model } = getAudioParam(context, beat, lang);
+    const { voiceId, provider, speechOptions, model } = MulmoStudioContextMethods.getAudioParam(context, beat, lang);
     const hash_string = [text, voiceId, speechOptions?.instruction ?? "", speechOptions?.speed ?? 1.0, provider, model ?? ""].join(":");
     GraphAILogger.log(`getBeatAudioPath [${hash_string}]`);
     const audioFileName = `${context.studio.filename}_${text2hash(hash_string)}`;
@@ -54,7 +47,7 @@ const preprocessorAgent = (namedInputs) => {
     const { beat, studioBeat, multiLingual, context, lang } = namedInputs;
     // const { lang } = context;
     const text = localizedText(beat, multiLingual, lang);
-    const { voiceId, provider, speechOptions, model } = getAudioParam(context, beat, lang);
+    const { voiceId, provider, speechOptions, model } = MulmoStudioContextMethods.getAudioParam(context, beat, lang);
     const audioPath = getBeatAudioPath(text, context, beat, lang);
     studioBeat.audioFile = audioPath; // TODO: Passing by reference is difficult to maintain, so pass it using graphai inputs
     const needsTTS = !beat.audio && audioPath !== undefined;

package/lib/agents/tts_google_agent.js

@@ -46,6 +46,6 @@ const ttsGoogleAgentInfo = {
     author: "Receptron Team",
     repository: "https://github.com/receptron/mulmocast-cli/",
     license: "MIT",
-    environmentVariables: ["GOOGLE_GENAI_API_KEY"],
+    environmentVariables: ["GEMINI_API_KEY"],
 };
 export default ttsGoogleAgentInfo;

package/lib/methods/mulmo_presentation_style.d.ts

@@ -26,10 +26,12 @@ export declare const MulmoPresentationStyleMethods: {
                 duration: number;
             } | undefined;
         };
+        keyName: string;
     };
     getSoundEffectAgentInfo(presentationStyle: MulmoPresentationStyle, beat: MulmoBeat): {
         agentName: string;
         defaultModel: import("../utils/provider2agent.js").ReplicateModel;
+        keyName: string;
         models: import("../utils/provider2agent.js").ReplicateModel[];
         modelParams: Record<import("../utils/provider2agent.js").ReplicateModel, {
             identifier?: `${string}/${string}:${string}`;
@@ -38,6 +40,7 @@ export declare const MulmoPresentationStyleMethods: {
     getLipSyncAgentInfo(presentationStyle: MulmoPresentationStyle, beat: MulmoBeat): {
         agentName: string;
         defaultModel: import("../utils/provider2agent.js").ReplicateModel;
+        keyName: string;
         models: import("../utils/provider2agent.js").ReplicateModel[];
         modelParams: Record<import("../utils/provider2agent.js").ReplicateModel, {
             identifier?: `${string}/${string}:${string}` | `${string}/${string}`;

package/lib/methods/mulmo_presentation_style.js

@@ -86,6 +86,7 @@ export const MulmoPresentationStyleMethods = {
         return {
             agent: agentInfo.agentName,
             imageParams: { ...defaultImageParams, ...imageParams },
+            keyName: agentInfo.keyName,
         };
     },
     getMovieAgentInfo(presentationStyle, beat) {
@@ -95,6 +96,7 @@ export const MulmoPresentationStyleMethods = {
         return {
             agent: agentInfo.agentName,
             movieParams,
+            keyName: agentInfo.keyName,
         };
     },
     getSoundEffectAgentInfo(presentationStyle, beat) {

package/lib/methods/mulmo_studio_context.d.ts

@@ -3,7 +3,7 @@
  * (No Node.js built-ins like fs, path, dotenv, etc.)
  * Works in both Node.js and modern browsers.
  */
-import { BeatSessionType, MulmoStudioContext, SessionProgressCallback, SessionType } from "../types/index.js";
+import { BeatSessionType, MulmoStudioContext, SessionProgressCallback, SessionType, MulmoBeat } from "../types/index.js";
 export declare const addSessionProgressCallback: (cb: SessionProgressCallback) => void;
 export declare const removeSessionProgressCallback: (cb: SessionProgressCallback) => void;
 export declare const MulmoStudioContextMethods: {
@@ -17,4 +17,13 @@ export declare const MulmoStudioContextMethods: {
     setBeatSessionState(context: MulmoStudioContext, sessionType: BeatSessionType | undefined, index: number, id: string | undefined, value: boolean): void;
     needTranslate(context: MulmoStudioContext, includeCaption?: boolean): boolean | "" | undefined;
     getIntroPadding(context: MulmoStudioContext): number;
+    getAudioParam(context: MulmoStudioContext, beat: MulmoBeat, lang?: string): {
+        voiceId: string;
+        provider: "google" | "mock" | "nijivoice" | "openai" | "elevenlabs";
+        speechOptions: {
+            speed?: number | undefined;
+            instruction?: string | undefined;
+        };
+        model: string | undefined;
+    };
 };

package/lib/methods/mulmo_studio_context.js

@@ -3,8 +3,10 @@
  * (No Node.js built-ins like fs, path, dotenv, etc.)
  * Works in both Node.js and modern browsers.
  */
+import { text2SpeechProviderSchema } from "../types/index.js";
 import { beatId } from "../utils/utils.js";
 import { GraphAILogger } from "graphai";
+import { MulmoPresentationStyleMethods } from "./mulmo_presentation_style.js";
 const sessionProgressCallbacks = new Set();
 export const addSessionProgressCallback = (cb) => {
     sessionProgressCallbacks.add(cb);
@@ -84,4 +86,10 @@ export const MulmoStudioContextMethods = {
         }
         return context.presentationStyle.audioParams.introPadding;
     },
+    getAudioParam(context, beat, lang) {
+        const speaker = MulmoPresentationStyleMethods.getSpeaker(context, beat, lang);
+        const speechOptions = { ...speaker.speechOptions, ...beat.speechOptions };
+        const provider = text2SpeechProviderSchema.parse(speaker.provider);
+        return { voiceId: speaker.voiceId, provider, speechOptions, model: speaker.model };
+    },
 };

package/lib/types/type.d.ts

@@ -76,6 +76,7 @@ export type PDFSize = (typeof pdf_sizes)[number];
 export type Text2ImageAgentInfo = {
     agent: string;
     imageParams: MulmoImageParams;
+    keyName?: string;
 };
 export type Text2HtmlAgentInfo = {
     provider: Text2HtmlImageProvider;

package/lib/utils/provider2agent.d.ts

@@ -2,22 +2,26 @@ export declare const provider2TTSAgent: {
     nijivoice: {
         agentName: string;
         hasLimitedConcurrency: boolean;
+        keyName: string;
     };
     openai: {
         agentName: string;
         hasLimitedConcurrency: boolean;
         defaultModel: string;
         defaultVoice: string;
+        keyName: string;
     };
     google: {
         agentName: string;
         hasLimitedConcurrency: boolean;
+        keyName: string;
     };
     elevenlabs: {
         agentName: string;
         hasLimitedConcurrency: boolean;
         defaultModel: string;
         models: string[];
+        keyName: string;
     };
     mock: {
         agentName: string;
@@ -31,21 +35,25 @@ export declare const provider2ImageAgent: {
         agentName: string;
         defaultModel: string;
         models: string[];
+        keyName: string;
     };
     google: {
         agentName: string;
         defaultModel: string;
         models: string[];
+        keyName: string;
     };
     replicate: {
         agentName: string;
         defaultModel: string;
         models: string[];
+        keyName: string;
     };
     mock: {
         agentName: string;
         defaultModel: string;
         models: string[];
+        keyName: string;
     };
 };
 export type ReplicateModel = `${string}/${string}`;
@@ -53,6 +61,7 @@ export declare const provider2MovieAgent: {
     replicate: {
         agentName: string;
         defaultModel: ReplicateModel;
+        keyName: string;
         models: string[];
         modelParams: Record<ReplicateModel, {
             durations: number[];
@@ -65,17 +74,20 @@ export declare const provider2MovieAgent: {
         agentName: string;
         defaultModel: string;
         models: string[];
+        keyName: string;
     };
     mock: {
         agentName: string;
         defaultModel: string;
         models: string[];
+        keyName: string;
     };
 };
 export declare const provider2SoundEffectAgent: {
     replicate: {
         agentName: string;
         defaultModel: ReplicateModel;
+        keyName: string;
         models: ReplicateModel[];
         modelParams: Record<ReplicateModel, {
             identifier?: `${string}/${string}:${string}`;
@@ -86,6 +98,7 @@ export declare const provider2LipSyncAgent: {
     replicate: {
         agentName: string;
         defaultModel: ReplicateModel;
+        keyName: string;
         models: ReplicateModel[];
         modelParams: Record<ReplicateModel, {
             identifier?: `${string}/${string}:${string}` | `${string}/${string}`;
@@ -99,6 +112,7 @@ export declare const provider2LLMAgent: {
     readonly openai: {
         readonly agentName: "openAIAgent";
         readonly defaultModel: "gpt-5";
+        readonly keyName: "OPENAI_API_KEY";
         readonly max_tokens: 8192;
         readonly models: readonly ["gpt-5", "gpt-5-nano", "gpt-5-mini", "gpt-4.1", "gpt-4.1-mini", "gpt-4.1-nano", "o3", "o3-mini", "o3-pro", "o1", "o1-pro", "gpt-4o", "gpt-4o-mini"];
     };
@@ -107,16 +121,19 @@ export declare const provider2LLMAgent: {
         readonly defaultModel: "claude-3-7-sonnet-20250219";
         readonly max_tokens: 8192;
         readonly models: readonly ["claude-opus-4-1-20250805", "claude-opus-4-20250514", "claude-sonnet-4-20250514", "claude-3-7-sonnet-20250219", "claude-3-haiku-20240307"];
+        readonly keyName: "ANTHROPIC_API_KEY";
     };
     readonly gemini: {
         readonly agentName: "geminiAgent";
         readonly defaultModel: "gemini-2.5-flash";
         readonly max_tokens: 8192;
         readonly models: readonly ["gemini-2.5-pro", "gemini-2.5-flash", "gemini-2.5-flash-lite", "gemini-2.0-flash"];
+        readonly keyName: "GEMINI_API_KEY";
     };
     readonly groq: {
         readonly agentName: "groqAgent";
         readonly defaultModel: "llama-3.1-8b-instant";
+        readonly keyName: "GROQ_API_KEY";
         readonly max_tokens: 4096;
         readonly models: readonly ["llama-3.1-8b-instant", "llama-3.3-70b-versatile", "deepseek-r1-distill-llama-70b", "openai/gpt-oss-120b", "openai/gpt-oss-20b"];
     };

package/lib/utils/provider2agent.js

@@ -3,16 +3,19 @@ export const provider2TTSAgent = {
     nijivoice: {
         agentName: "ttsNijivoiceAgent",
         hasLimitedConcurrency: true,
+        keyName: "NIJIVOICE_API_KEY",
     },
     openai: {
         agentName: "ttsOpenaiAgent",
         hasLimitedConcurrency: false,
         defaultModel: "gpt-4o-mini-tts",
         defaultVoice: "shimmer",
+        keyName: "OPENAI_API_KEY",
     },
     google: {
         agentName: "ttsGoogleAgent",
         hasLimitedConcurrency: false,
+        keyName: "GEMINI_API_KEY",
     },
     elevenlabs: {
         agentName: "ttsElevenlabsAgent",
@@ -21,6 +24,7 @@ export const provider2TTSAgent = {
         // Models | ElevenLabs Documentation
         // https://elevenlabs.io/docs/models
         models: ["eleven_multilingual_v2", "eleven_turbo_v2_5", "eleven_turbo_v2", "eleven_flash_v2_5", "eleven_flash_v2"],
+        keyName: "ELEVENLABS_API_KEY",
     },
     mock: {
         agentName: "mediaMockAgent",
@@ -34,27 +38,32 @@ export const provider2ImageAgent = {
         agentName: "imageOpenaiAgent",
         defaultModel: "gpt-image-1",
         models: ["dall-e-3", "gpt-image-1"],
+        keyName: "OPENAI_API_KEY",
     },
     google: {
         agentName: "imageGenAIAgent",
         defaultModel: "gemini-2.5-flash-image-preview",
         models: ["imagen-3.0-generate-002", "imagen-4.0-generate-preview-06-06", "imagen-4.0-ultra-generate-preview-06-06", "gemini-2.5-flash-image-preview"],
+        keyName: "GEMINI_API_KEY",
     },
     replicate: {
         agentName: "imageReplicateAgent",
         defaultModel: "bytedance/seedream-4",
         models: ["bytedance/seedream-4", "qwen/qwen-image"],
+        keyName: "REPLICATE_API_TOKEN",
     },
     mock: {
         agentName: "mediaMockAgent",
         defaultModel: "mock-model",
         models: ["mock-model"],
+        keyName: "",
     },
 };
 export const provider2MovieAgent = {
     replicate: {
         agentName: "movieReplicateAgent",
         defaultModel: "bytedance/seedance-1-lite",
+        keyName: "REPLICATE_API_TOKEN",
         models: [
             "bytedance/seedance-1-lite",
             "bytedance/seedance-1-pro",
@@ -151,17 +160,20 @@ export const provider2MovieAgent = {
         agentName: "movieGenAIAgent",
         defaultModel: "veo-2.0-generate-001",
         models: ["veo-2.0-generate-001", "veo-3.0-generate-preview"],
+        keyName: "GEMINI_API_KEY",
     },
     mock: {
         agentName: "mediaMockAgent",
         defaultModel: "mock-model",
         models: ["mock-model"],
+        keyName: "",
     },
 };
 export const provider2SoundEffectAgent = {
     replicate: {
         agentName: "soundEffectReplicateAgent",
         defaultModel: "zsxkib/mmaudio",
+        keyName: "REPLICATE_API_TOKEN",
         models: ["zsxkib/mmaudio"],
         modelParams: {
             "zsxkib/mmaudio": {
@@ -174,6 +186,7 @@ export const provider2LipSyncAgent = {
     replicate: {
         agentName: "lipSyncReplicateAgent",
         defaultModel: "bytedance/omni-human",
+        keyName: "REPLICATE_API_TOKEN",
         models: ["bytedance/latentsync", "tmappdev/lipsync", "bytedance/omni-human"],
         modelParams: {
             "bytedance/latentsync": {
@@ -212,6 +225,7 @@ export const provider2LLMAgent = {
     openai: {
         agentName: "openAIAgent",
         defaultModel: "gpt-5",
+        keyName: "OPENAI_API_KEY",
         max_tokens: 8192,
         models: [
             "gpt-5",
@@ -234,16 +248,19 @@ export const provider2LLMAgent = {
         defaultModel: "claude-3-7-sonnet-20250219",
         max_tokens: 8192,
         models: ["claude-opus-4-1-20250805", "claude-opus-4-20250514", "claude-sonnet-4-20250514", "claude-3-7-sonnet-20250219", "claude-3-haiku-20240307"],
+        keyName: "ANTHROPIC_API_KEY",
     },
     gemini: {
         agentName: "geminiAgent",
         defaultModel: "gemini-2.5-flash",
         max_tokens: 8192,
         models: ["gemini-2.5-pro", "gemini-2.5-flash", "gemini-2.5-flash-lite", "gemini-2.0-flash"],
+        keyName: "GEMINI_API_KEY",
     },
     groq: {
         agentName: "groqAgent",
         defaultModel: "llama-3.1-8b-instant",
+        keyName: "GROQ_API_KEY",
         max_tokens: 4096,
         models: ["llama-3.1-8b-instant", "llama-3.3-70b-versatile", "deepseek-r1-distill-llama-70b", "openai/gpt-oss-120b", "openai/gpt-oss-20b"],
     },

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "mulmocast",
-  "version": "1.2.34",
+  "version": "1.2.36",
   "description": "",
   "type": "module",
   "main": "lib/index.node.js",
@@ -18,7 +18,8 @@
     }
   },
   "bin": {
-    "mulmo": "lib/cli/bin.js"
+    "mulmo": "lib/cli/bin.js",
+    "mulmo-mcp": "lib/mcp/server.js"
   },
   "files": [
     "./lib",
@@ -69,7 +70,7 @@
   "homepage": "https://github.com/receptron/mulmocast-cli#readme",
   "dependencies": {
     "@google-cloud/text-to-speech": "^6.3.0",
-    "@google/genai": "^1.19.0",
+    "@google/genai": "^1.20.0",
     "@graphai/anthropic_agent": "^2.0.11",
     "@graphai/browserless_agent": "^2.0.1",
     "@graphai/gemini_agent": "^2.0.1",
@@ -90,7 +91,7 @@
     "graphai": "^2.0.15",
     "jsdom": "^27.0.0",
     "marked": "^16.3.0",
-    "mulmocast-vision": "^1.0.3",
+    "mulmocast-vision": "^1.0.4",
     "ora": "^8.2.0",
     "puppeteer": "^24.20.0",
     "replicate": "^1.1.0",
@@ -111,7 +112,7 @@
     "prettier": "^3.6.2",
     "tsx": "^4.20.5",
     "typescript": "^5.9.2",
-    "typescript-eslint": "^8.43.0"
+    "typescript-eslint": "^8.44.0"
   },
   "engines": {
     "node": ">=18.0.0"