cerevox 3.0.0-beta.26 → 3.0.0-beta.28

package/dist/mcp/servers/zerocut.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"zerocut.d.ts","sourceRoot":"","sources":["../../../src/mcp/servers/zerocut.ts"],"names":[],"mappings":";AA64KA,wBAAsB,GAAG,kBAKxB"}
+{"version":3,"file":"zerocut.d.ts","sourceRoot":"","sources":["../../../src/mcp/servers/zerocut.ts"],"names":[],"mappings":";AAk4KA,wBAAsB,GAAG,kBAKxB"}

package/dist/mcp/servers/zerocut.js

@@ -956,7 +956,7 @@ server.registerTool('upload-custom-material', {
 });
 server.registerTool('generate-image', {
     title: 'Generate Image',
-    description: `生成图片`,
+    description: `生成图片，支持批量生成1-4张图`,
     inputSchema: {
         type: zod_1.z
             .enum([
@@ -1025,9 +1025,16 @@ server.registerTool('generate-image', {
         ])
             .default('720x1280')
             .describe('The size of the image.'),
-        saveToFileName: zod_1.z
-            .string()
-            .describe('The filename to save. 应该是png文件'),
+        imageCount: zod_1.z
+            .number()
+            .min(1)
+            .max(4)
+            .optional()
+            .default(1)
+            .describe('The number of images to generate. 暂时最多支持4张图，多了容易超时'),
+        saveToFileNames: zod_1.z
+            .array(zod_1.z.string())
+            .describe('The filenames to save. 数量要和imageCount对应，应该是png文件'),
         watermark: zod_1.z
             .boolean()
             .optional()
@@ -1061,7 +1068,7 @@ server.registerTool('generate-image', {
 \`\`\`
           `),
     },
-}, async ({ type = 'seedream', prompt, sceneIndex, storyBoardFile = 'storyboard.json', skipConsistencyCheck = false, size = '720x1280', saveToFileName, watermark, referenceImages, }) => {
+}, async ({ type = 'seedream', prompt, sceneIndex, storyBoardFile = 'storyboard.json', skipConsistencyCheck = false, size = '720x1280', imageCount = 1, saveToFileNames, watermark, referenceImages, }) => {
     try {
         // 验证session状态
         const currentSession = await validateSession('generate-image');
@@ -1071,7 +1078,6 @@ server.registerTool('generate-image', {
             checkStoryboardFlag = true;
             return createErrorResponse('必须先审查生成的 storyboard.json 内容，确保每个场景中的stage_atmosphere内容按照规则被正确融合到start_frame和video_prompt中，不得遗漏，检查完成后先汇报，如果有问题，应当先修改 storyboard.json 内容，然后再调用 generate-image 生成图片。注意修改 storyboard 内容时，仅修改相应字段的字符串值，不要破坏JSON格式！', 'generate-image');
         }
-        const validatedFileName = validateFileName(saveToFileName);
         // 校验 prompt 与 storyboard.json 中场景设定的一致性
         if (sceneIndex && !skipConsistencyCheck) {
             try {
@@ -1220,6 +1226,9 @@ ${referenceImages?.map((ref, index) => `图${index + 1}：${ref.image}`).join('\
         catch (error) {
             console.error('Failed to optimize prompt:', error);
         }
+        if (imageCount > 1) {
+            processedPrompt = `$请生成${imageCount}张相关图片 ${processedPrompt}`;
+        }
         console.log(`Generating image with prompt: ${processedPrompt.substring(0, 100)}...`);
         // 处理参考图片
         let imageBase64Array;
@@ -1301,13 +1310,25 @@ ${processedPrompt}`.trim();
         if (!res) {
             throw new Error('Failed to generate image: no response from AI service');
         }
-        if (res.url) {
+        if (res.urls && res.urls.length > 0) {
             console.log('Image generated successfully, saving to materials...');
-            const uri = await saveMaterial(currentSession, res.url, validatedFileName);
+            let uris = [];
+            if (res.urls.length === 1) {
+                uris = [
+                    await saveMaterial(currentSession, res.urls[0], validateFileName(saveToFileNames[0])),
+                ];
+            }
+            else {
+                // 多图场景
+                uris = await Promise.all(res.urls.map((url, i) => {
+                    const fileName = validateFileName(saveToFileNames[i]);
+                    return saveMaterial(currentSession, url, fileName);
+                }));
+            }
             const result = {
                 success: true,
                 // source: res.url,
-                uri,
+                uris,
                 prompt: processedPrompt,
                 size,
                 timestamp: new Date().toISOString(),
@@ -2364,7 +2385,7 @@ server.registerTool('generate-scene-tts', {
             .describe('The volume of the tts.'),
         voiceID: zod_1.z
             .string()
-            .describe(`适合作为视频配音的音色ID，除非用户指定，否则你必须已通过 search_voice 工具检查确定该音色确实是存在的。`),
+            .describe(`适合作为视频配音的音色ID，除非用户指定，否则你必须确保已通过 pick-voice 工具挑选出真实存在的音色。`),
         context_texts: zod_1.z
             .array(zod_1.z.string())
             .default([])
@@ -2398,7 +2419,7 @@ server.registerTool('generate-scene-tts', {
             try {
                 const voice = (await ai.listVoices()).find(v => v.id === voiceID);
                 if (!voice) {
-                    return createErrorResponse(`Voice ${voiceID} not found in voice-list. Use search-voices tool to find available voices. 若用户坚持要使用该音色，需跳过一致性检查。`, 'generate-scene-tts');
+                    return createErrorResponse(`Voice ${voiceID} not found in voice-list. Use pick-voice tool to pick an available voice. 若用户坚持要使用该音色，需跳过一致性检查。`, 'generate-scene-tts');
                 }
                 const storyBoardPath = (0, node_path_1.resolve)(process.env.ZEROCUT_PROJECT_CWD || process.cwd(), projectLocalDir, storyBoardFile);
                 if ((0, node_fs_1.existsSync)(storyBoardPath)) {
@@ -2451,7 +2472,7 @@ server.registerTool('generate-scene-tts', {
         }
         console.log(`Generating TTS with voice: ${voiceID}, speed: ${finalSpeed}, text: ${text.substring(0, 100)}...`);
         if (voiceID.startsWith('BV0')) {
-            throw new Error(`BV0* 系列音色已弃用，你必须已通过 search_voice 工具检查确定该音色确实是存在的。`);
+            throw new Error(`BV0* 系列音色已弃用，你必须通过 pick-voice 工具挑选一个真实存在的音色。`);
         }
         const type = voiceID.startsWith('zh_') ||
             voiceID.startsWith('en_') ||
@@ -2606,7 +2627,7 @@ ${context_texts.join('\n')}
                         type: 'text',
                         text: JSON.stringify({
                             success: false,
-                            error: 'No TTS URL returned from AI service. You should use search-voices tool to find available voices.',
+                            error: 'No TTS URL returned from AI service. You should use pick-voice tool to pick an available voice.',
                             response: res,
                             timestamp: new Date().toISOString(),
                         }),
@@ -2839,67 +2860,71 @@ server.registerTool('get-schema', {
         return createErrorResponse(error, 'get-schema');
     }
 });
-server.registerTool('search-voices', {
-    title: 'Search Voices',
-    description: 'Search voices from doubao_voices_full based on languages, and gender. 搜索并选择符合要求的语音，在合适的情况下，优先采用 volcano_tts_2 类型的语音',
+server.registerTool('pick-voice', {
+    title: 'Pick Voice',
+    description: '根据用户需求，选择尽可能符合要求的语音，在合适的情况下，优先采用 volcano_tts_2 类型的语音',
     inputSchema: {
-        languages: zod_1.z
-            .array(zod_1.z.enum([
-            'zh',
-            'en',
-            'ja',
-            'ko',
-            'es',
-            'pt',
-            'nl',
-            'vi',
-            'ru',
-            'id',
-            'de',
-            'fr',
-            'it',
-            'ar',
-            'tr',
-            'uk',
-        ]))
+        prompt: zod_1.z
+            .string()
+            .describe('用户需求描述，例如：一个有亲和力的，适合给孩子讲故事的语音'),
+        custom_design: zod_1.z
+            .boolean()
             .optional()
-            .describe('Filter by languages (e.g., ["zh", "en"]). If not provided, no language filtering is applied.'),
-        gender: zod_1.z
-            .enum(['male', 'female'])
+            .describe('是否自定义语音，由于要消耗较多积分，因此**只有用户明确要求自己设计语音**，才将该参数设为true'),
+        custom_design_preview: zod_1.z
+            .string()
             .optional()
-            .describe('Filter by gender (male or female). If not provided, no gender filtering is applied.'),
+            .describe('用户自定义语音的预览文本，用于展示自定义语音的效果，只有 custom_design 为 true 时才需要'),
+        custom_design_save_to: zod_1.z
+            .string()
+            .optional()
+            .describe('自定义语音的保存路径，例如：custom_voice.mp3 custom_voice_{id}.mp3'),
     },
-}, async ({ languages, gender }) => {
+}, async ({ prompt, custom_design, custom_design_preview, custom_design_save_to, }) => {
     try {
         // 验证session状态
-        const currentSession = await validateSession('search-voices');
+        const currentSession = await validateSession('pick-voice');
         const ai = currentSession.ai;
-        let filteredVoices = await ai.listVoices();
-        // Filter by languages
-        if (languages && languages.length > 0) {
-            filteredVoices = filteredVoices.filter(voice => voice.languages &&
-                voice.languages.some((lang) => languages.includes(lang)));
-        }
-        // Filter by gender
-        if (gender) {
-            filteredVoices = filteredVoices.filter(voice => {
-                return voice.gender === gender;
+        if (custom_design) {
+            if (!custom_design_preview) {
+                throw new Error('custom_design_preview is required when custom_design is true');
+            }
+            const data = await currentSession.ai.voiceDesign({
+                prompt,
+                previewText: custom_design_preview,
             });
+            if (data.voice_id) {
+                const trial_audio = data.trial_audio;
+                let uri = '';
+                if (trial_audio) {
+                    uri = await saveMaterial(currentSession, trial_audio, custom_design_save_to || `custom_voice_${data.voice_id}.mp3`);
+                }
+                return {
+                    content: [
+                        {
+                            type: 'text',
+                            text: JSON.stringify({
+                                success: true,
+                                ...data,
+                                uri,
+                                timestamp: new Date().toISOString(),
+                            }),
+                        },
+                    ],
+                };
+            }
+            else {
+                throw new Error(`Voice design failed, ${JSON.stringify(data)}`);
+            }
         }
+        const data = await ai.pickVoice({ prompt });
         return {
             content: [
                 {
                     type: 'text',
                     text: JSON.stringify({
                         success: true,
-                        data: {
-                            totalCount: filteredVoices.length,
-                            voices: filteredVoices,
-                            filters: {
-                                languages: languages || null,
-                                gender: gender || null,
-                            },
-                        },
+                        ...data,
                         timestamp: new Date().toISOString(),
                     }),
                 },
@@ -2907,52 +2932,7 @@ server.registerTool('search-voices', {
         };
     }
     catch (error) {
-        return createErrorResponse(error, 'search-voices');
-    }
-});
-server.registerTool('voice-design', {
-    title: 'Voice Design',
-    description: 'Design a voice based on a prompt. The voice will be designed based on the prompt and preview text.',
-    inputSchema: {
-        prompt: zod_1.z.string().describe('The prompt to design the voice.'),
-        previewText: zod_1.z.string().describe('The preview text to design the voice.'),
-        saveToFileName: zod_1.z
-            .string()
-            .describe('The file name to save the designed voice. 应该是mp3文件'),
-    },
-}, async ({ prompt, previewText, saveToFileName }) => {
-    try {
-        const currentSession = await validateSession('voice-design');
-        const data = await currentSession.ai.voiceDesign({
-            prompt,
-            previewText,
-        });
-        if (data.voice_id) {
-            const trial_audio = data.trial_audio;
-            let uri = '';
-            if (trial_audio) {
-                uri = await saveMaterial(currentSession, trial_audio, saveToFileName);
-            }
-            return {
-                content: [
-                    {
-                        type: 'text',
-                        text: JSON.stringify({
-                            success: true,
-                            ...data,
-                            uri,
-                            timestamp: new Date().toISOString(),
-                        }),
-                    },
-                ],
-            };
-        }
-        else {
-            throw new Error(`Voice design failed, ${JSON.stringify(data)}`);
-        }
-    }
-    catch (error) {
-        return createErrorResponse(error, 'voice-design');
+        return createErrorResponse(error, 'pick-voice');
     }
 });
 server.registerTool('media-analyzer', {