mulmocast 2.6.11 → 2.6.12

package/README.md

@@ -139,7 +139,7 @@ BROWSERLESS_API_TOKEN=your_browserless_api_token # to access web in mulmo tool
 
 ### Google Vertex AI
 
-For enterprise/production environments or to access models like Imagen 4, use Vertex AI with Application Default Credentials (ADC):
+For enterprise/production environments, or to use Veo movie models that are only available on Vertex AI, use Vertex AI with Application Default Credentials (ADC):
 
 ```bash
 # Install gcloud CLI and authenticate
@@ -151,7 +151,7 @@ Configure in MulmoScript:
 {
   "imageParams": {
     "provider": "google",
-    "model": "imagen-4.0-generate-001",
+    "model": "gemini-2.5-flash-image",
     "vertexai_project": "your-project-id",
     "vertexai_location": "us-central1"
   }

package/lib/agents/image_genai_agent.d.ts

@@ -1,5 +1,6 @@
 import type { AgentFunction, AgentFunctionInfo } from "graphai";
 import type { AgentBufferResult, ImageAgentInputs, ImageAgentParams, GenAIImageAgentConfig } from "../types/agent.js";
+export declare const buildDeprecatedGoogleImageModelMessage: (model: string) => string | null;
 export declare const imageGenAIAgent: AgentFunction<ImageAgentParams, AgentBufferResult, ImageAgentInputs, GenAIImageAgentConfig>;
 declare const imageGenAIAgentInfo: AgentFunctionInfo;
 export default imageGenAIAgentInfo;

package/lib/agents/image_genai_agent.js

@@ -1,10 +1,16 @@
 import fs from "fs";
 import { GraphAILogger } from "graphai";
-import { provider2ImageAgent } from "../types/provider2agent.js";
-import { apiKeyMissingError, agentIncorrectAPIKeyError, agentGenerationError, agentInvalidResponseError, imageAction, imageFileTarget, hasCause, getGenAIErrorReason, resultify, } from "../utils/error_cause.js";
+import { provider2ImageAgent, deprecatedGoogleImageModelHints, vertexAIGlobalOnlyImageModels, } from "../types/provider2agent.js";
+import { apiKeyMissingError, agentIncorrectAPIKeyError, agentGenerationError, agentInvalidResponseError, imageAction, imageFileTarget, unsupportedModelTarget, hasCause, getGenAIErrorReason, resultify, } from "../utils/error_cause.js";
 import { getAspectRatio } from "../utils/utils.js";
 import { ASPECT_RATIOS, PRO_ASPECT_RATIOS } from "../types/const.js";
 import { GoogleGenAI, PersonGeneration } from "@google/genai";
+const isDeprecatedGoogleImageModel = (model) => model in deprecatedGoogleImageModelHints;
+export const buildDeprecatedGoogleImageModelMessage = (model) => {
+    if (!isDeprecatedGoogleImageModel(model))
+        return null;
+    return `Google image model "${model}" is no longer available. ${deprecatedGoogleImageModelHints[model]}`;
+};
 const getGeminiContents = (prompt, referenceImages) => {
     const contents = [{ text: prompt }];
     const images = [...(referenceImages ?? [])];
@@ -58,11 +64,17 @@ const errorProcess = (error) => {
 export const imageGenAIAgent = async ({ namedInputs, params, config, }) => {
     const { prompt, referenceImages } = namedInputs;
     const model = params.model ?? provider2ImageAgent["google"].defaultModel;
+    const deprecatedMessage = buildDeprecatedGoogleImageModelMessage(model);
+    if (deprecatedMessage) {
+        throw new Error(deprecatedMessage, {
+            cause: agentGenerationError("imageGenAIAgent", imageAction, unsupportedModelTarget),
+        });
+    }
     const apiKey = config?.apiKey;
     const ai = params.vertexai_project
         ? (() => {
             const location = params.vertexai_location ?? "us-central1";
-            if (model === "gemini-3-pro-image-preview" && location !== "global") {
+            if (vertexAIGlobalOnlyImageModels.has(model) && location !== "global") {
                 GraphAILogger.warn(`imageGenAIAgent: model "${model}" on Vertex AI is only available in location "global", but got "${location}". Set imageParams.vertexai_location to "global".`);
             }
             return new GoogleGenAI({ vertexai: true, project: params.vertexai_project, location });

package/lib/types/provider2agent.d.ts

@@ -47,6 +47,16 @@ export declare const deprecatedOpenAIImageModelHints: {
     readonly "dall-e-3": "Use 'gpt-image-1' or another supported model.";
 };
 export type DeprecatedOpenAIImageModel = keyof typeof deprecatedOpenAIImageModelHints;
+export declare const deprecatedGoogleImageModelHints: {
+    readonly "imagen-3.0-generate-002": "Use 'gemini-2.5-flash-image' or 'gemini-3-pro-image-preview' instead.";
+    readonly "imagen-4.0-generate-001": "Use 'gemini-2.5-flash-image' or 'gemini-3-pro-image-preview' instead.";
+    readonly "imagen-4.0-ultra-generate-001": "Use 'gemini-2.5-flash-image' or 'gemini-3-pro-image-preview' instead.";
+    readonly "imagen-4.0-fast-generate-001": "Use 'gemini-2.5-flash-image' or 'gemini-3-pro-image-preview' instead.";
+    readonly "imagen-4.0-generate-preview-06-06": "Use 'gemini-2.5-flash-image' or 'gemini-3-pro-image-preview' instead.";
+    readonly "imagen-4.0-ultra-generate-preview-06-06": "Use 'gemini-2.5-flash-image' or 'gemini-3-pro-image-preview' instead.";
+};
+export type DeprecatedGoogleImageModel = keyof typeof deprecatedGoogleImageModelHints;
+export declare const vertexAIGlobalOnlyImageModels: ReadonlySet<string>;
 export declare const provider2ImageAgent: {
     openai: {
         agentName: string;

package/lib/types/provider2agent.js

@@ -50,6 +50,20 @@ export const deprecatedOpenAIImageModelHints = {
     "dall-e-2": supportedOpenAIImageReplacementHint,
     "dall-e-3": supportedOpenAIImageReplacementHint,
 };
+const supportedGoogleImageReplacementHint = "Use 'gemini-2.5-flash-image' or 'gemini-3-pro-image-preview' instead.";
+export const deprecatedGoogleImageModelHints = {
+    "imagen-3.0-generate-002": supportedGoogleImageReplacementHint,
+    "imagen-4.0-generate-001": supportedGoogleImageReplacementHint,
+    "imagen-4.0-ultra-generate-001": supportedGoogleImageReplacementHint,
+    "imagen-4.0-fast-generate-001": supportedGoogleImageReplacementHint,
+    "imagen-4.0-generate-preview-06-06": supportedGoogleImageReplacementHint,
+    "imagen-4.0-ultra-generate-preview-06-06": supportedGoogleImageReplacementHint,
+};
+// Google image models that on Vertex AI are only published under location "global"
+// (regional endpoints like us-central1 return 404 NOT_FOUND).
+// See https://cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/3-pro-image
+// and https://cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/3-1-flash-image
+export const vertexAIGlobalOnlyImageModels = new Set(["gemini-3-pro-image-preview", "gemini-3.1-flash-image-preview"]);
 export const provider2ImageAgent = {
     openai: {
         agentName: "imageOpenaiAgent",
@@ -61,14 +75,7 @@ export const provider2ImageAgent = {
     google: {
         agentName: "imageGenAIAgent",
         defaultModel: "gemini-2.5-flash-image",
-        models: [
-            "imagen-4.0-generate-001",
-            "imagen-4.0-ultra-generate-001",
-            "imagen-4.0-fast-generate-001",
-            "gemini-2.5-flash-image",
-            "gemini-3.1-flash-image-preview",
-            "gemini-3-pro-image-preview",
-        ],
+        models: ["gemini-2.5-flash-image", "gemini-3.1-flash-image-preview", "gemini-3-pro-image-preview"],
         keyName: "GEMINI_API_KEY",
     },
     replicate: {

package/lib/utils/ffmpeg_utils.d.ts

@@ -1,6 +1,7 @@
-import ffmpeg from "fluent-ffmpeg";
+import ffmpeg from "@modernized/fluent-ffmpeg";
+type FfmpegCommand = ReturnType<typeof ffmpeg>;
 export type FfmpegContext = {
-    command: ffmpeg.FfmpegCommand;
+    command: FfmpegCommand;
     inputCount: number;
     filterComplex: string[];
 };
@@ -29,3 +30,4 @@ export declare const extractImageFromMovie: (movieFile: string, imagePath: strin
 export declare const trimMusic: (inputFile: string, startTime: number, duration: number) => Promise<Buffer>;
 export declare const createSilentAudio: (filePath: string, durationSec: number) => Promise<void>;
 export declare const pcmToMp3: (rawPcm: Buffer, sampleRate?: number) => Promise<Buffer>;
+export {};

package/lib/utils/ffmpeg_utils.js

@@ -1,4 +1,4 @@
-import ffmpeg from "fluent-ffmpeg";
+import ffmpeg from "@modernized/fluent-ffmpeg";
 import { GraphAILogger } from "graphai";
 import { isFile } from "./file.js";
 import fs from "fs";

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "mulmocast",
-  "version": "2.6.11",
+  "version": "2.6.12",
   "description": "",
   "type": "module",
   "main": "lib/index.node.js",
@@ -89,7 +89,7 @@
   "homepage": "https://github.com/receptron/mulmocast-cli#readme",
   "dependencies": {
     "@google-cloud/text-to-speech": "^6.4.1",
-    "@google/genai": "^1.50.1",
+    "@google/genai": "^2.0.1",
     "@graphai/anthropic_agent": "^2.0.12",
     "@graphai/browserless_agent": "^2.0.2",
     "@graphai/gemini_agent": "^2.0.5",
@@ -102,12 +102,12 @@
     "@inquirer/input": "^5.0.12",
     "@inquirer/select": "^5.1.4",
     "@modelcontextprotocol/sdk": "^1.29.0",
+    "@modernized/fluent-ffmpeg": "^0.1.4",
     "@mozilla/readability": "^0.6.0",
-    "@tavily/core": "^0.5.11",
+    "@tavily/core": "^0.7.3",
     "archiver": "^7.0.1",
     "clipboardy": "^5.3.1",
     "dotenv": "^17.4.2",
-    "fluent-ffmpeg": "^2.1.3",
     "graphai": "^2.0.16",
     "jsdom": "^29.1.1",
     "marked": "^18.0.3",
@@ -123,7 +123,6 @@
     "@eslint/js": "^10.0.1",
     "@receptron/test_utils": "^2.0.3",
     "@types/archiver": "^7.0.0",
-    "@types/fluent-ffmpeg": "^2.1.28",
     "@types/jsdom": "^28.0.1",
     "@types/yargs": "^17.0.35",
     "cross-env": "^10.1.0",

package/scripts/test/test_genai.json

@@ -34,16 +34,11 @@
       }
     },
     {
-      "id": "imagen_4",
-      "text": "image generated by imagen-4",
-      "imagePrompt": "a woman is walking through a busy Tokyo street at night, she is wearing dark sunglasses"
-    },
-    {
-      "id": "imagen_4_ultra",
-      "text": "image generated by imagen-4",
-      "imagePrompt": "a woman is walking through a busy Tokyo street at night, she is wearing dark sunglasses",
+      "id": "gemini_2_5_flash_image_alt",
+      "text": "image generated by gemini-2.5-flash-image with a different prompt (used by image_animation_showcase to give the carousel/grid four distinct images)",
+      "imagePrompt": "a beautiful mountain landscape at sunrise with snow-capped peaks",
       "imageParams": {
-        "model": "imagen-4.0-ultra-generate-preview-06-06"
+        "model": "gemini-2.5-flash-image"
       }
     },
     {

package/scripts/test/test_images_imagen_deprecated.json

@@ -0,0 +1,55 @@
+{
+  "$mulmocast": {
+    "version": "1.1"
+  },
+  "lang": "en",
+  "title": "Test Deprecated Google Imagen Models",
+  "imageParams": {
+    "provider": "google",
+    "style": "Photorealistic-style"
+  },
+  "beats": [
+    {
+      "text": "imagen-3.0-generate-002 is deprecated; mulmocast rejects this with a migration hint before calling the Gemini API",
+      "imagePrompt": "Blue sky, a flock of birds",
+      "imageParams": {
+        "model": "imagen-3.0-generate-002"
+      }
+    },
+    {
+      "text": "imagen-4.0-generate-001 is deprecated; mulmocast rejects this with a migration hint before calling the Gemini API",
+      "imagePrompt": "Blue sky, a flock of birds",
+      "imageParams": {
+        "model": "imagen-4.0-generate-001"
+      }
+    },
+    {
+      "text": "imagen-4.0-ultra-generate-001 is deprecated; mulmocast rejects this with a migration hint before calling the Gemini API",
+      "imagePrompt": "Blue sky, a flock of birds",
+      "imageParams": {
+        "model": "imagen-4.0-ultra-generate-001"
+      }
+    },
+    {
+      "text": "imagen-4.0-fast-generate-001 is deprecated; mulmocast rejects this with a migration hint before calling the Gemini API",
+      "imagePrompt": "Blue sky, a flock of birds",
+      "imageParams": {
+        "model": "imagen-4.0-fast-generate-001"
+      }
+    },
+    {
+      "text": "imagen-4.0-generate-preview-06-06 is deprecated; mulmocast rejects this with a migration hint before calling the Gemini API",
+      "imagePrompt": "Blue sky, a flock of birds",
+      "imageParams": {
+        "model": "imagen-4.0-generate-preview-06-06"
+      }
+    },
+    {
+      "text": "imagen-4.0-ultra-generate-preview-06-06 is deprecated; mulmocast rejects this with a migration hint before calling the Gemini API",
+      "imagePrompt": "Blue sky, a flock of birds",
+      "imageParams": {
+        "model": "imagen-4.0-ultra-generate-preview-06-06"
+      }
+    }
+  ]
+}

package/scripts/test/test_vertexai.json

@@ -1,7 +1,7 @@
 {
   "$mulmocast": { "version": "1.1" },
   "lang": "en",
-  "title": "Vertex AI Test (TTS + Image + Movie)",
+  "title": "Vertex AI Test (TTS + Image + Movie) — image variants",
   "description": "Replace YOUR_PROJECT_ID with your Google Cloud project ID",
   "speechParams": {
     "speakers": {
@@ -18,7 +18,7 @@
   },
   "imageParams": {
     "provider": "google",
-    "model": "imagen-4.0-generate-001",
+    "model": "gemini-2.5-flash-image",
     "vertexai_project": "YOUR_PROJECT_ID",
     "vertexai_location": "us-central1"
   },
@@ -30,37 +30,37 @@
   },
   "beats": [
     {
-      "id": "gcp_tts_imagen4",
+      "id": "gcp_tts_gemini_flash_image",
       "speaker": "GCP",
-      "text": "TTS: Google Cloud TTS en-US-Studio-O. Image: Imagen 4.",
+      "text": "TTS: Google Cloud TTS en-US-Studio-O. Image: Gemini 2.5 Flash Image.",
       "imagePrompt": "A beautiful mountain landscape at sunrise"
     },
     {
-      "id": "gemini_tts_imagen4",
+      "id": "gemini_tts_gemini_flash_image",
       "speaker": "Gemini",
-      "text": "TTS: Gemini 2.5 Pro TTS Kore. Image: Imagen 4.",
+      "text": "TTS: Gemini 2.5 Pro TTS Kore. Image: Gemini 2.5 Flash Image.",
       "imagePrompt": "A futuristic city with flying cars"
     },
     {
-      "id": "imagen4_ultra_override",
+      "id": "gemini_pro_image_override",
       "speaker": "GCP",
-      "text": "TTS: Google Cloud TTS en-US-Studio-O. Image: Imagen 4 Ultra override.",
+      "text": "TTS: Google Cloud TTS en-US-Studio-O. Image: Gemini 3 Pro Image override.",
       "imagePrompt": "A woman walking through Tokyo at night with neon lights",
       "imageParams": {
-        "model": "imagen-4.0-ultra-generate-001",
+        "model": "gemini-3-pro-image-preview",
         "vertexai_project": "YOUR_PROJECT_ID",
-        "vertexai_location": "us-central1"
+        "vertexai_location": "global"
       }
     },
     {
-      "id": "imagen4_fast_override",
+      "id": "gemini_3_1_flash_image_override",
       "speaker": "Gemini",
-      "text": "TTS: Gemini 2.5 Pro TTS Kore. Image: Imagen 4 Fast override.",
+      "text": "TTS: Gemini 2.5 Pro TTS Kore. Image: Gemini 3.1 Flash Image override.",
       "imagePrompt": "A cat sitting on a windowsill watching rain",
       "imageParams": {
-        "model": "imagen-4.0-fast-generate-001",
+        "model": "gemini-3.1-flash-image-preview",
         "vertexai_project": "YOUR_PROJECT_ID",
-        "vertexai_location": "us-central1"
+        "vertexai_location": "global"
       }
     },
     {
@@ -84,7 +84,7 @@
     {
       "id": "image_to_movie_veo2",
       "speaker": "GCP",
-      "text": "TTS: Google Cloud TTS en-US-Studio-O. Image: Imagen 4. Movie: Veo 2 with image reference.",
+      "text": "TTS: Google Cloud TTS en-US-Studio-O. Image: Gemini 2.5 Flash Image. Movie: Veo 2 with image reference.",
       "imagePrompt": "A woman standing in a busy Tokyo street at night",
       "moviePrompt": "The woman takes a selfie with her phone",
       "duration": 5
@@ -92,12 +92,12 @@
     {
       "id": "image_to_movie_veo3_override",
       "speaker": "Gemini",
-      "text": "TTS: Gemini 2.5 Pro TTS Kore. Image: Imagen 4 Fast override. Movie: Veo 3 override with image reference.",
+      "text": "TTS: Gemini 2.5 Pro TTS Kore. Image: Gemini 3.1 Flash Image override. Movie: Veo 3 override with image reference.",
       "imagePrompt": "A robot standing in a futuristic laboratory",
       "imageParams": {
-        "model": "imagen-4.0-fast-generate-001",
+        "model": "gemini-3.1-flash-image-preview",
         "vertexai_project": "YOUR_PROJECT_ID",
-        "vertexai_location": "us-central1"
+        "vertexai_location": "global"
       },
       "moviePrompt": "The robot turns its head and waves at the camera",
       "movieParams": {

package/scripts/test/test_vertexai_simple.json

@@ -5,7 +5,7 @@
   "description": "Replace YOUR_PROJECT_ID with your Google Cloud project ID",
   "imageParams": {
     "provider": "google",
-    "model": "imagen-4.0-generate-001",
+    "model": "gemini-2.5-flash-image",
     "vertexai_project": "YOUR_PROJECT_ID"
   },
   "beats": [