cerevox 3.0.0-beta.9 → 3.0.2

package/dist/mcp/servers/zerocut.js

@@ -50,7 +50,6 @@ const constants_1 = require("../../utils/constants");
 const videokit_1 = require("../../utils/videokit");
 const promises_1 = require("node:fs/promises");
 const node_path_1 = __importStar(require("node:path"));
-const doubao_voices_full_1 = require("./helper/doubao_voices_full");
 const node_fs_1 = require("node:fs");
 const coze_1 = require("../../utils/coze");
 const mp3_duration_1 = __importDefault(require("mp3-duration"));
@@ -94,10 +93,13 @@ function validateFileName(fileName) {
     if (!fileName || fileName.trim() === '') {
         throw new Error('File name cannot be empty');
     }
+    if (fileName.startsWith('materials/')) {
+        fileName = fileName.replace('materials/', '');
+    }
     if (fileName.includes('..') ||
         fileName.includes('/') ||
         fileName.includes('\\')) {
-        throw new Error('Invalid file name: contains illegal characters');
+        throw new Error('Invalid file name: contains illegal characters, cannot contain ".." or "/" or "\\"');
     }
     return fileName.trim();
 }
@@ -130,6 +132,7 @@ async function initProject(session) {
     return workDir;
 }
 async function saveMaterial(session, url, saveToFileName) {
+    saveToFileName = validateFileName(saveToFileName);
     const terminal = session.terminal;
     const saveToPath = `/home/user/cerevox-zerocut/projects/${terminal.id}/materials/${saveToFileName}`;
     const saveLocalPath = (0, node_path_1.resolve)(projectLocalDir, 'materials', saveToFileName);
@@ -514,7 +517,7 @@ server.registerTool('project-close', {
             .min(0)
             .max(20)
             .default(5)
-            .describe('Close the session after the specified number of minutes. Default is 5 minutes. 如果用户要求立即关闭会话，请将该参数设置为0！'),
+            .describe('Close the session after the specified number of minutes. Default is 5 minutes. 除非用户要求立即关闭会话，将该参数设置为0，否则应默认设为5'),
     },
 }, async ({ inMinutes }) => {
     try {
@@ -663,69 +666,6 @@ server.registerTool('retrieve-rules-context', {
         return createErrorResponse(`Failed to load rules context prompt for ${prompt}: ${error}`, 'retrieve-rules-context');
     }
 });
-// 列出项目下的所有文件
-server.registerTool('list-project-files', {
-    title: 'List Project Files',
-    description: 'List all files in the materials directory.',
-    inputSchema: {},
-}, async () => {
-    try {
-        // 验证session状态
-        const currentSession = await validateSession('list-project-files');
-        console.log('Listing project files...');
-        const terminal = currentSession.terminal;
-        if (!terminal) {
-            throw new Error('Terminal not available in current session');
-        }
-        let cwd;
-        try {
-            cwd = await terminal.getCwd();
-        }
-        catch (cwdError) {
-            console.error('Failed to get current working directory:', cwdError);
-            throw new Error('Failed to get current working directory');
-        }
-        console.log(`Current working directory: ${cwd}`);
-        // 安全地列出各目录文件，失败时返回空数组
-        const listFilesWithFallback = async (path, dirName) => {
-            try {
-                const files = await currentSession.files.listFiles(path);
-                console.log(`Found ${files?.length || 0} files in ${dirName}`);
-                return files || [];
-            }
-            catch (error) {
-                console.warn(`Failed to list files in ${dirName} (${path}):`, error);
-                return [];
-            }
-        };
-        const [rootFiles, materialsFiles, outputFiles] = await Promise.all([
-            listFilesWithFallback(cwd, 'root'),
-            listFilesWithFallback(`${cwd}/materials`, 'materials'),
-            listFilesWithFallback(`${cwd}/output`, 'output'),
-        ]);
-        const result = {
-            success: true,
-            cwd,
-            root: rootFiles,
-            materials: materialsFiles,
-            output: outputFiles,
-            totalFiles: rootFiles.length + materialsFiles.length + outputFiles.length,
-            timestamp: new Date().toISOString(),
-        };
-        console.log(`Total files found: ${result.totalFiles}`);
-        return {
-            content: [
-                {
-                    type: 'text',
-                    text: JSON.stringify(result),
-                },
-            ],
-        };
-    }
-    catch (error) {
-        return createErrorResponse(error, 'list-project-files');
-    }
-});
 server.registerTool('generate-character-image', {
     title: 'Generate Character Image',
     description: 'Generate a turnaround image or portrait for any character.',
@@ -932,49 +872,6 @@ ${roleDescriptionPrompt}
         return createErrorResponse(error, 'generate-character-image');
     }
 });
-server.registerTool('generate-line-sketch', {
-    title: 'Generate Line Sketch',
-    description: 'Generate line sketch material based on user prompt.',
-    inputSchema: {
-        prompt: zod_1.z.string().describe('The prompt to generate line sketch.'),
-        saveToFileName: zod_1.z
-            .string()
-            .describe('The filename to save the generated line sketch. 应该是png文件'),
-    },
-}, async ({ prompt, saveToFileName }) => {
-    try {
-        // 验证session状态
-        await validateSession('generate-line-sketch');
-        // 验证文件名
-        validateFileName(saveToFileName);
-        // 调用AI生成线稿
-        const res = await session.ai.generateLineSketch({ prompt });
-        if (res && res.url) {
-            // 保存到本地
-            await saveMaterial(session, res.url, saveToFileName);
-            const result = {
-                success: true,
-                url: res.url,
-                localPath: getMaterialUri(session, saveToFileName),
-                timestamp: new Date().toISOString(),
-            };
-            return {
-                content: [
-                    {
-                        type: 'text',
-                        text: JSON.stringify(result),
-                    },
-                ],
-            };
-        }
-        else {
-            throw new Error('No URL returned from AI service');
-        }
-    }
-    catch (error) {
-        return createErrorResponse(error, 'generate-line-sketch');
-    }
-});
 server.registerTool('upload-custom-material', {
     title: 'Upload Custom Material',
     description: 'Upload material files (images: jpeg/png, videos: mp4, audio: mp3) from the local filesystem to the materials directory. For video and audio files, duration information will be included in the response.',
@@ -1063,15 +960,21 @@ 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1-9张图，若用户要求生成关联图片或组图，请一次生成，不要分几次生成`,
     inputSchema: {
         type: zod_1.z
-            .enum(['banana', 'banana-pro', 'seedream', 'seedream-pro'])
+            .enum([
+            'banana',
+            'banana-pro',
+            'seedream',
+            'seedream-pro',
+            'line-sketch',
+        ])
             .optional()
             .default('seedream'),
         prompt: zod_1.z
             .string()
-            .describe('The prompt to generate. 一般要严格对应 storyboard 中当前场景的 start_frame 或 end_frame 中的字段描述'),
+            .describe('The prompt to generate. 一般要严格对应 storyboard 中当前场景的 start_frame 或 end_frame 中的字段描述，如果是生成线稿，则 type 使用 line-sketch'),
         sceneIndex: zod_1.z
             .number()
             .min(1)
@@ -1126,9 +1029,16 @@ 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(9)
+            .optional()
+            .default(1)
+            .describe('The number of images to generate. 最多支持9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()
@@ -1162,7 +1072,7 @@ 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, }, context) => {
     try {
         // 验证session状态
         const currentSession = await validateSession('generate-image');
@@ -1172,7 +1082,6 @@ 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 {
@@ -1321,6 +1230,9 @@ ${referenceImages?.map((ref, index) => `图${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（每张都是大图，非拼接） ${processedPrompt}`;
+        }
         console.log(`Generating image with prompt: ${processedPrompt.substring(0, 100)}...`);
         // 处理参考图片
         let imageBase64Array;
@@ -1392,23 +1304,55 @@ ${processedPrompt}`.trim();
             }
         }
         const ai = currentSession.ai;
-        const res = await ai.generateImage({
+        const { taskUrl } = await ai.generateImage({
             type,
             prompt: processedPrompt,
             size,
             watermark,
             image: imageBase64Array,
+            async: true,
+        });
+        let progress = 0;
+        const res = await ai.waitForTaskComplete({
+            taskUrl,
+            onProgress: async (metaData) => {
+                try {
+                    await sendProgress(context, ++progress, undefined, JSON.stringify(metaData));
+                }
+                catch (progressError) {
+                    console.warn('Failed to send progress update:', progressError);
+                }
+            },
         });
         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 && res.urls[0]) {
+                uris = [
+                    await saveMaterial(currentSession, res.urls[0], validateFileName(saveToFileNames[0])),
+                ];
+            }
+            else {
+                // 多图场景
+                uris = await Promise.all(res.urls.map((url, i) => {
+                    if (!url)
+                        return '';
+                    try {
+                        const fileName = validateFileName(saveToFileNames[i]);
+                        return saveMaterial(currentSession, url, fileName);
+                    }
+                    catch (ex) {
+                        return '';
+                    }
+                }));
+            }
             const result = {
                 success: true,
                 // source: res.url,
-                uri,
+                uris,
                 prompt: processedPrompt,
                 size,
                 timestamp: new Date().toISOString(),
@@ -1579,7 +1523,7 @@ server.registerTool('edit-image', {
 let lastEffect = '';
 server.registerTool('generate-video', {
     title: 'Generate Video',
-    description: `生成视频`,
+    description: `图生视频和首尾帧生视频工具`,
     inputSchema: {
         prompt: zod_1.z
             .string()
@@ -1611,6 +1555,8 @@ server.registerTool('generate-video', {
             'hailuo-fast',
             'vidu',
             'vidu-pro',
+            'vidu-uc',
+            'vidu-uc-pro',
             'kling',
             'kling-pro',
             'pixv',
@@ -2029,7 +1975,7 @@ server.registerTool('generate-video', {
                     console.warn('Failed to send progress update:', progressError);
                 }
             },
-            waitForFinish: true,
+            waitForFinish: type !== 'zero',
         });
         if (!res) {
             throw new Error('Failed to generate video: no response from AI service');
@@ -2077,7 +2023,7 @@ server.registerTool('generate-video', {
                         type: 'text',
                         text: JSON.stringify({
                             success: true,
-                            message: '该视频生成任务正在运行中，它是异步任务，且执行时间较长，你应立即调用工具 wait-for-task-finish 来等待任务结束，如该工具调用超时，你应立即再次重新调用直到任务结束。',
+                            message: '该视频生成任务正在运行中，它是异步任务，且执行时间较长，你应立即调用工具 wait-for-task-finish 来等待任务结束，如 wait-for-task-finish 工具调用超时，你应立即再次重新调用直到任务结束。',
                             taskUrl: res.taskUrl,
                         }),
                     },
@@ -2107,7 +2053,7 @@ server.registerTool('generate-video', {
     }
 });
 server.registerTool('wait-for-task-finish', {
-    title: 'Wait Workflow or VideoTask Done',
+    title: 'Wait Workflow or VideoTask Done；只有正在运行Coze工作流或者有异步生成视频任务时才需要执行这个工具',
     description: 'Wait for a workflow to complete.',
     inputSchema: {
         taskUrl: zod_1.z
@@ -2122,7 +2068,7 @@ server.registerTool('wait-for-task-finish', {
         const currentSession = await validateSession('wait-for-task-finish');
         const ai = currentSession.ai;
         let progress = 0;
-        const res = await ai.waitForVideoTaskComplete({
+        const res = await ai.waitForTaskComplete({
             taskUrl,
             onProgress: async (metaData) => {
                 try {
@@ -2237,14 +2183,32 @@ server.registerTool('generate-sound-effect', {
         return createErrorResponse(error, 'generate-sound-effect');
     }
 });
-server.registerTool('generate-music', {
-    title: 'Generate Music',
-    description: 'Generate the music. Include background music or song.',
+server.registerTool('generate-music-or-mv', {
+    title: '创作音乐（Music）或音乐视频（Music Video）',
+    description: '生成音乐，包括MV（music video）、BGM 或 歌曲',
     inputSchema: {
         prompt: zod_1.z.string().describe('The prompt to generate.'),
+        singerPhoto: zod_1.z
+            .string()
+            .optional()
+            .describe('The singer photo to use. 只有type为music_video的时候才生效，也可以不传，模型会自动生成'),
+        mvOrientation: zod_1.z
+            .enum(['portrait', 'landscape'])
+            .optional()
+            .describe('The orientation of the music video. Defaults to portrait.')
+            .default('portrait'),
+        mvOriginalSong: zod_1.z
+            .string()
+            .optional()
+            .describe('用于生成mv的音乐. 只有type为music_video的时候才生效，也可以不传，模型会自动创作'),
+        mvGenSubtitles: zod_1.z
+            .boolean()
+            .optional()
+            .default(false)
+            .describe('是否生成mv的字幕. 默认为false，只有type为music_video的时候才生效'),
         type: zod_1.z
-            .enum(['bgm', 'song'])
-            .describe('The type of music. Defaults to background music.')
+            .enum(['bgm', 'song', 'music_video'])
+            .describe('The type of music. Defaults to BGM. ⚠️ 如果 type 是 music_video，会直接生成音频和视频，**不需要**额外专门生成歌曲')
             .default('bgm'),
         model: zod_1.z
             .enum(['doubao', 'minimax'])
@@ -2262,9 +2226,9 @@ server.registerTool('generate-music', {
             .describe('Whether to skip copyright check.'),
         saveToFileName: zod_1.z
             .string()
-            .describe('The filename to save. 应该是mp3文件'),
+            .describe('The filename to save. 如果type是music video，应该是mp4文件，否则应该是mp3文件'),
     },
-}, async ({ prompt, type, model, duration, skipCopyCheck, saveToFileName }, context) => {
+}, async ({ prompt, singerPhoto, mvOrientation, mvOriginalSong, mvGenSubtitles, type, model, duration, skipCopyCheck, saveToFileName, }, context) => {
     try {
         // 验证session状态
         const currentSession = await validateSession('generate-music');
@@ -2275,19 +2239,49 @@ server.registerTool('generate-music', {
         if (type === 'bgm' && duration > 120) {
             throw new Error('BGM duration must be at most 120 seconds.');
         }
-        const finalPrompt = `${prompt.trim()} ${type === 'bgm' ? `纯音乐无歌词，时长${duration}秒` : `时长${duration}秒，使用${model}模型`}`;
-        const res = await ai.generateMusic({
-            prompt: finalPrompt,
-            skipCopyCheck,
-            onProgress: async (metaData) => {
-                try {
-                    await sendProgress(context, metaData.Result?.Progress ?? ++progress, metaData.Result?.Progress ? 100 : undefined, JSON.stringify(metaData));
-                }
-                catch (progressError) {
-                    console.warn('Failed to send progress update:', progressError);
-                }
-            },
-        });
+        let res;
+        if (type === 'music_video') {
+            const singer_photo = singerPhoto
+                ? await getMaterialUri(currentSession, singerPhoto)
+                : undefined;
+            const original_song = mvOriginalSong
+                ? await getMaterialUri(currentSession, mvOriginalSong)
+                : undefined;
+            res = await ai.generateZeroCutMusicVideo({
+                // prompt: `${prompt.trim()} 音乐时长${duration}秒`,
+                prompt,
+                singerPhoto: singer_photo,
+                orientation: mvOrientation,
+                genSubtitles: mvGenSubtitles,
+                originalSong: original_song,
+                duration,
+                resolution: '720p',
+                onProgress: async (metaData) => {
+                    try {
+                        await sendProgress(context, metaData.Result?.Progress ?? ++progress, metaData.Result?.Progress ? 100 : undefined, JSON.stringify(metaData));
+                    }
+                    catch (progressError) {
+                        console.warn('Failed to send progress update:', progressError);
+                    }
+                },
+                waitForFinish: false,
+            });
+        }
+        else {
+            const finalPrompt = `${prompt.trim()} ${type === 'bgm' ? `纯音乐无歌词，时长${duration}秒` : `时长${duration}秒，使用${model}模型`}`;
+            res = await ai.generateMusic({
+                prompt: finalPrompt,
+                skipCopyCheck,
+                onProgress: async (metaData) => {
+                    try {
+                        await sendProgress(context, metaData.Result?.Progress ?? ++progress, metaData.Result?.Progress ? 100 : undefined, JSON.stringify(metaData));
+                    }
+                    catch (progressError) {
+                        console.warn('Failed to send progress update:', progressError);
+                    }
+                },
+            });
+        }
         if (!res) {
             throw new Error('Failed to generate Music: no response from AI service');
         }
@@ -2330,6 +2324,20 @@ server.registerTool('generate-music', {
                 ],
             };
         }
+        else if (res.taskUrl) {
+            return {
+                content: [
+                    {
+                        type: 'text',
+                        text: JSON.stringify({
+                            success: true,
+                            message: '该视频生成任务正在运行中，它是异步任务，且执行时间较长，你应立即调用工具 wait-for-task-finish 来等待任务结束，如如 wait-for-task-finish 工具调用超时，你应立即再次重新调用直到任务结束。',
+                            taskUrl: res.taskUrl,
+                        }),
+                    },
+                ],
+            };
+        }
         else {
             console.warn('Music generation completed but no URL returned');
             return {
@@ -2399,44 +2407,44 @@ server.registerTool('generate-scene-tts', {
             .optional()
             .default(1.0)
             .describe('The volume of the tts.'),
-        emotion: zod_1.z
-            .enum([
-            'storytelling',
-            'neutral',
-            'excited',
-            'coldness',
-            'angry',
-            'sad',
-            'happy',
-            'surprised',
-            'fear',
-            'depressed',
-            'lovey-dovey',
-            'shy',
-            'comfort',
-            'tension',
-            'tender',
-            'magnetic',
-            'vocal - fry',
-            'ASMR',
-        ])
-            .optional(),
         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([])
+            .describe(`语音合成的辅助信息，用于模型对话式合成，能更好的体现语音情感
+
+可以探索，比如常见示例有以下几种：
+
+1. 语速调整
+  - context_texts: ["你可以说慢一点吗？"]
+2. 情绪/语气调整
+  - context_texts=["你可以用特别特别痛心的语气说话吗?"]
+  - context_texts=["嗯，你的语气再欢乐一点"]
+3. 音量调整
+  - context_texts=["你嗓门再小点。"]
+4. 音感调整
+  - context_texts=["你能用骄傲的语气来说话吗？"]
+`),
         explicit_language: zod_1.z.enum(['zh', 'en', 'ja']).optional().default('zh'),
     },
-}, async ({ text, sceneIndex, storyBoardFile, skipConsistencyCheck, voiceID, saveToFileName, speed, pitch, volume, emotion, explicit_language, }) => {
+}, async ({ text, sceneIndex, storyBoardFile, skipConsistencyCheck, voiceID, saveToFileName, speed, pitch, volume, context_texts, explicit_language, }) => {
     try {
         // 验证session状态
         const currentSession = await validateSession('generate-scene-tts');
         const validatedFileName = validateFileName(saveToFileName);
         const finalSpeed = speed ?? 1;
         volume = volume ?? 1;
+        const ai = currentSession.ai;
         let scene = null;
         // 校验 text 与 storyboard.json 中场景设定的一致性
         if (sceneIndex && !skipConsistencyCheck) {
             try {
+                const voice = (await ai.listVoices()).find(v => v.id === voiceID);
+                if (!voice) {
+                    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)) {
                     const storyBoardContent = await (0, promises_1.readFile)(storyBoardPath, 'utf8');
@@ -2487,9 +2495,8 @@ server.registerTool('generate-scene-tts', {
             }
         }
         console.log(`Generating TTS with voice: ${voiceID}, speed: ${finalSpeed}, text: ${text.substring(0, 100)}...`);
-        const ai = currentSession.ai;
         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_') ||
@@ -2499,30 +2506,80 @@ server.registerTool('generate-scene-tts', {
             ? 'volcano'
             : 'minimax';
         let res;
+        let emotion = 'auto';
         if (type === 'volcano') {
-            const voice = doubao_voices_full_1.doubaoVoicesFull.find(v => v.voiceID === voiceID);
-            if (!voice) {
-                return createErrorResponse(`Voice ${voiceID} not found in Doubao voices. Use search-voices tool to find available voices.`, 'generate-scene-tts');
-            }
-            const emotions = voice.emotions || [];
-            if (emotion && !emotions.includes(emotion)) {
-                emotion = 'neutral';
-            }
-            // 修复可能的 emotion 错误情况
-            emotion = emotion || 'neutral';
             volume = Math.max(Math.min(volume, 2.0), 0.5);
             res = await ai.textToSpeechVolc({
                 text: text.trim(),
                 speaker: voiceID,
                 speed: Math.floor(100 * (finalSpeed - 1)),
                 volume: Math.floor(100 * (volume - 1)),
-                emotion,
+                context_texts,
                 explicit_language,
                 voice_to_caption: explicit_language === 'zh' || explicit_language === 'en',
             });
         }
         else {
-            emotion = emotion || 'neutral';
+            emotion = 'neutral';
+            if (context_texts.length > 0) {
+                const prompt = `根据用户输入语音内容和上下文内容，从文字判断语音合理的情感，然后选择以下情感**之一**返回结果：
+
+"happy", "sad", "angry", "fearful", "disgusted", "surprised", "calm", "fluent", "whisper", "neutral"
+
+## 要求
+输出 JSON 格式，包含一个 emotion 字段，值为以上情感之一。
+`;
+                const schema = {
+                    name: 'emotion_schema',
+                    schema: {
+                        type: 'object',
+                        properties: {
+                            emotion: {
+                                type: 'string',
+                                enum: [
+                                    'neutral',
+                                    'happy',
+                                    'sad',
+                                    'angry',
+                                    'fearful',
+                                    'disgusted',
+                                    'surprised',
+                                    'calm',
+                                    'fluent',
+                                    'whisper',
+                                ],
+                                description: '用户输入语音的情感',
+                            },
+                        },
+                        required: ['emotion'],
+                    },
+                };
+                const payload = {
+                    model: 'Doubao-Seed-1.6',
+                    messages: [
+                        {
+                            role: 'system',
+                            content: prompt,
+                        },
+                        {
+                            role: 'user',
+                            content: `## 语音内容：
+${text.trim()}
+
+## 语音上下文
+${context_texts.join('\n')}
+`,
+                        },
+                    ],
+                    response_format: {
+                        type: 'json_schema',
+                        json_schema: schema,
+                    },
+                };
+                const completion = await ai.getCompletions(payload);
+                const emotionObj = JSON.parse(completion.choices[0]?.message?.content ?? '{}');
+                emotion = emotionObj.emotion ?? 'neutral';
+            }
             res = await ai.textToSpeech({
                 text: text.trim(),
                 voiceName: voiceID,
@@ -2563,6 +2620,8 @@ server.registerTool('generate-scene-tts', {
                 uri,
                 durationMs: Math.floor((duration || 0) * 1000),
                 text,
+                emotion,
+                context_texts,
                 voiceName: voiceID,
                 speed: finalSpeed,
                 timestamp: new Date().toISOString(),
@@ -2592,7 +2651,7 @@ server.registerTool('generate-scene-tts', {
                         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(),
                         }),
@@ -2825,111 +2884,71 @@ 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 scenes, emotions, languages, and gender.',
+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: {
-        scenes: zod_1.z
-            .array(zod_1.z.enum([
-            'asmr',
-            'audiobook',
-            'customer_service',
-            'dialect_fun',
-            'dialogue',
-            'kids_content',
-            'news_explainer',
-            'podcast_voiceover',
-            'product_ad',
-            'promo_trailer',
-            'roleplay_drama',
-            'story_narration',
-            'storytelling',
-            'tutorial',
-        ]))
-            .optional()
-            .describe('Filter by scenes (e.g., ["product_ad", "tutorial"]). If not provided, no scene filtering is applied.'),
-        emotions: zod_1.z
-            .array(zod_1.z.enum([
-            'ASMR',
-            'affectionate',
-            'angry',
-            'authoritative',
-            'chat',
-            'coldness',
-            'depressed',
-            'excited',
-            'fear',
-            'happy',
-            'hate',
-            'neutral',
-            'sad',
-            'surprised',
-            'warm',
-        ]))
+        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 emotions (e.g., ["happy", "neutral"]). If not provided, no emotion filtering is applied.'),
-        languages: zod_1.z
-            .array(zod_1.z.string())
+            .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 languages (e.g., ["zh", "en-US"]). If not provided, no language filtering is applied.'),
-        gender: zod_1.z
-            .enum(['male', 'female'])
+            .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('Filter by gender (male or female). If not provided, no gender filtering is applied.'),
+            .describe('自定义语音的保存路径，例如：custom_voice.mp3 custom_voice_{id}.mp3'),
     },
-}, async ({ scenes, emotions, languages, gender }) => {
+}, async ({ prompt, custom_design, custom_design_preview, custom_design_save_to, }) => {
     try {
-        let filteredVoices = [...doubao_voices_full_1.doubaoVoicesFull];
-        // Filter by scenes
-        if (scenes && scenes.length > 0) {
-            filteredVoices = filteredVoices.filter(voice => voice.scenes &&
-                voice.scenes.some(scene => scenes.includes(scene)));
-        }
-        // Filter by emotions
-        if (emotions && emotions.length > 0) {
-            filteredVoices = filteredVoices.filter(voice => {
-                // If emotions includes 'neutral', also include voices without emotions field
-                if (emotions.includes('neutral') && !voice.emotions) {
-                    return true;
-                }
-                return (voice.emotions &&
-                    voice.emotions.some(emotion => emotions.includes(emotion)));
+        // 验证session状态
+        const currentSession = await validateSession('pick-voice');
+        const ai = currentSession.ai;
+        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,
             });
-        }
-        // 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 => {
-                const voiceId = voice.voiceID.toLowerCase();
-                if (gender === 'male') {
-                    return voiceId.includes('_male_');
+            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`);
                 }
-                else if (gender === 'female') {
-                    return voiceId.includes('_female_');
-                }
-                return true;
-            });
+                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: {
-                                scenes: scenes || null,
-                                emotions: emotions || null,
-                                languages: languages || null,
-                                gender: gender || null,
-                            },
-                        },
+                        ...data,
                         timestamp: new Date().toISOString(),
                     }),
                 },
@@ -2937,52 +2956,7 @@ 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', {
@@ -3198,206 +3172,6 @@ server.registerTool('media-analyzer', {
         return createErrorResponse(error, 'media-analyzer');
     }
 });
-// server.registerTool(
-//   'image-aligner',
-//   {
-//     title: 'Image Aligner',
-//     description:
-//       'Analyze image quality and alignment with prompt using AI Image Quality Inspector.',
-//     inputSchema: {
-//       imageFileName: z
-//         .string()
-//         .describe('The image file name in materials directory to analyze.'),
-//       sceneIndex: z.number().min(1).describe('场景索引，从1开始的下标'),
-//       storyBoardFile: z
-//         .string()
-//         .optional()
-//         .default('storyboard.json')
-//         .describe('故事板文件路径'),
-//       imagePrompt: z
-//         .string()
-//         .optional()
-//         .describe('可选的图片提示词，如果提供则覆盖storyboard中的提示词'),
-//       customPrompt: z
-//         .string()
-//         .optional()
-//         .describe('可选的额外用户要求，用于补充图片质量评估的特定需求'),
-//     },
-//   },
-//   async ({
-//     imageFileName,
-//     sceneIndex,
-//     storyBoardFile = 'storyboard.json',
-//     imagePrompt,
-//     customPrompt,
-//   }) => {
-//     try {
-//       const currentSession = await validateSession('image-aligner');
-//       // 验证图片文件
-//       validateImageFile(imageFileName);
-//       // 获取图片 URL
-//       const imageUrl = getMaterialUri(currentSession, imageFileName);
-//       // 确定要使用的提示词
-//       let finalPrompt = imagePrompt;
-//       // 如果没有提供imagePrompt，则从storyboard中获取
-//       if (!imagePrompt) {
-//         try {
-//           const storyBoardPath = resolve(
-//             process.env.ZEROCUT_PROJECT_CWD || process.cwd(),
-//             projectLocalDir,
-//             storyBoardFile
-//           );
-//           if (existsSync(storyBoardPath)) {
-//             const storyBoardContent = await readFile(storyBoardPath, 'utf8');
-//             const storyBoard = JSON.parse(storyBoardContent);
-//             if (storyBoard.scenes && Array.isArray(storyBoard.scenes)) {
-//               const scene = storyBoard.scenes[sceneIndex - 1]; // sceneIndex 从1开始，数组从0开始
-//               if (scene) {
-//                 // 根据文件名判断优先级：若end_frame存在且imageFileName包含"_end"则优先取end_frame，否则取start_frame
-//                 if (scene.end_frame && imageFileName.includes('_end')) {
-//                   finalPrompt = scene.end_frame;
-//                 } else {
-//                   finalPrompt = scene.start_frame || scene.end_frame;
-//                 }
-//                 if (!finalPrompt) {
-//                   return createErrorResponse(
-//                     `场景 ${sceneIndex} 中未找到 start_frame 或 end_frame 提示词`,
-//                     'image-aligner'
-//                   );
-//                 }
-//               } else {
-//                 return createErrorResponse(
-//                   `在 ${storyBoardFile} 中未找到场景索引 ${sceneIndex}`,
-//                   'image-aligner'
-//                 );
-//               }
-//             } else {
-//               return createErrorResponse(
-//                 `${storyBoardFile} 文件格式不正确，缺少 scenes 数组`,
-//                 'image-aligner'
-//               );
-//             }
-//           } else {
-//             return createErrorResponse(
-//               `故事板文件不存在: ${storyBoardPath}`,
-//               'image-aligner'
-//             );
-//           }
-//         } catch (error) {
-//           return createErrorResponse(
-//             `读取或解析故事板文件失败: ${error}`,
-//             'image-aligner'
-//           );
-//         }
-//       }
-//       // 如果仍然没有提示词，返回错误
-//       if (!finalPrompt) {
-//         return createErrorResponse(
-//           '未提供 imagePrompt 且无法从故事板中获取提示词',
-//           'image-aligner'
-//         );
-//       }
-//       // 读取图片质量检查指南
-//       const alignerGuidelinePath = resolve(
-//         __dirname,
-//         './prompts/reasonings/image_aligner.md'
-//       );
-//       let alignerGuideline = '';
-//       try {
-//         alignerGuideline = await readFile(alignerGuidelinePath, 'utf8');
-//       } catch (error) {
-//         console.warn('无法读取图片质量检查指南:', error);
-//         alignerGuideline =
-//           '请对图片质量进行评估，包括构图、色彩、清晰度等方面。';
-//       }
-//       // 构建系统提示
-//       const systemPrompt = `你是一个专业的AI图片质量检查员。请根据以下指南对图片进行评估：
-// ${alignerGuideline}
-// 请严格按照指南中的JSON格式返回评估结果。`;
-//       // 构建用户提示
-//       const userPrompt = `请对这张图片进行质量评估。
-// 原始提示词：${finalPrompt}${
-//         customPrompt
-//           ? `
-// 额外要求：${customPrompt}`
-//           : ''
-//       }
-// 请按照指南要求，返回包含评分、问题列表和优化建议的JSON格式结果。`;
-//       // 调用AI模型进行图片质量评估
-//       const ai = currentSession.ai;
-//       const completion = await ai.getCompletions({
-//         model: 'Doubao-Seed-1.6',
-//         messages: [
-//           {
-//             role: 'system',
-//             content: systemPrompt,
-//           },
-//           {
-//             role: 'user',
-//             content: [
-//               {
-//                 type: 'image_url',
-//                 image_url: {
-//                   url: imageUrl,
-//                 },
-//               },
-//               {
-//                 type: 'text',
-//                 text: userPrompt,
-//               },
-//             ],
-//           },
-//         ],
-//       });
-//       const result = completion.choices[0]?.message?.content;
-//       if (!result) {
-//         throw new Error('No response from AI model');
-//       }
-//       // 解析AI响应
-//       let alignmentResult;
-//       try {
-//         // 尝试从响应中提取JSON
-//         const jsonMatch =
-//           result.match(/```json\s*([\s\S]*?)\s*```/) ||
-//           result.match(/\{[\s\S]*\}/);
-//         if (jsonMatch) {
-//           alignmentResult = JSON.parse(jsonMatch[1] || jsonMatch[0]);
-//         } else {
-//           // 如果没有找到JSON格式，尝试直接解析整个响应
-//           alignmentResult = JSON.parse(result);
-//         }
-//       } catch (error) {
-//         // 如果解析失败，返回原始响应
-//         alignmentResult = {
-//           error: 'JSON解析失败',
-//           raw_response: result,
-//         };
-//       }
-//       return {
-//         content: [
-//           {
-//             type: 'text',
-//             text: JSON.stringify({
-//               success: true,
-//               imageFileName,
-//               sceneIndex,
-//               storyBoardFile,
-//               imagePrompt: finalPrompt,
-//               customPrompt,
-//               promptSource: imagePrompt ? 'manual_override' : 'storyboard',
-//               analysis: alignmentResult,
-//               imageUrl,
-//               nextActionSuggest:
-//                 '可根据分析结果调整提示词，修改storyboard后，重新生成图片。',
-//             }),
-//           },
-//         ],
-//       };
-//     } catch (error) {
-//       return createErrorResponse(error, 'image-aligner');
-//     }
-//   }
-// );
 server.registerTool('audio-video-sync', {
     title: 'Audio Video Sync',
     description: 'Generate audio-video-synced video by matching video with audio. 还可以对口型。',
@@ -3552,7 +3326,7 @@ server.registerTool('audio-video-sync', {
     }
 });
 server.registerTool('generate-video-by-ref', {
-    title: 'Generate Video by Reference Images',
+    title: '参考生视频（包含文生视频）工具',
     description: 'Generate video using reference images. Supports sora2, sora2-pro (1 image max), veo3.1, veo3.1-pro (1 image max), lite and pro (4 images max), vidu (7 images max). Can work without reference images (0 images).',
     inputSchema: {
         prompt: zod_1.z
@@ -3602,6 +3376,7 @@ server.registerTool('generate-video-by-ref', {
             'veo3.1',
             'veo3.1-pro',
             'vidu',
+            'vidu-uc',
             'pixv',
         ])
             .default('lite')
@@ -3895,7 +3670,7 @@ ${prompt.trim()}`,
                         type: 'text',
                         text: JSON.stringify({
                             success: true,
-                            message: '该视频生成任务正在运行中，它是异步任务，且执行时间较长，你应立即调用工具 wait-for-task-finish 来等待任务结束，如该工具调用超时，你应立即再次重新调用直到任务结束。',
+                            message: '该视频生成任务正在运行中，它是异步任务，且执行时间较长，你应立即调用工具 wait-for-task-finish 来等待任务结束，如如 wait-for-task-finish 工具调用超时，你应立即再次重新调用直到任务结束。',
                             taskUrl: result.taskUrl,
                         }),
                     },
@@ -4328,6 +4103,69 @@ server.registerTool('search-context', {
         return createErrorResponse(error, 'search-context');
     }
 });
+// 列出项目下的所有文件
+server.registerTool('list-project-files', {
+    title: 'List Project Files',
+    description: 'List all files in the materials directory.',
+    inputSchema: {},
+}, async () => {
+    try {
+        // 验证session状态
+        const currentSession = await validateSession('list-project-files');
+        console.log('Listing project files...');
+        const terminal = currentSession.terminal;
+        if (!terminal) {
+            throw new Error('Terminal not available in current session');
+        }
+        let cwd;
+        try {
+            cwd = await terminal.getCwd();
+        }
+        catch (cwdError) {
+            console.error('Failed to get current working directory:', cwdError);
+            throw new Error('Failed to get current working directory');
+        }
+        console.log(`Current working directory: ${cwd}`);
+        // 安全地列出各目录文件，失败时返回空数组
+        const listFilesWithFallback = async (path, dirName) => {
+            try {
+                const files = await currentSession.files.listFiles(path);
+                console.log(`Found ${files?.length || 0} files in ${dirName}`);
+                return files || [];
+            }
+            catch (error) {
+                console.warn(`Failed to list files in ${dirName} (${path}):`, error);
+                return [];
+            }
+        };
+        const [rootFiles, materialsFiles, outputFiles] = await Promise.all([
+            listFilesWithFallback(cwd, 'root'),
+            listFilesWithFallback(`${cwd}/materials`, 'materials'),
+            listFilesWithFallback(`${cwd}/output`, 'output'),
+        ]);
+        const result = {
+            success: true,
+            cwd,
+            root: rootFiles,
+            materials: materialsFiles,
+            output: outputFiles,
+            totalFiles: rootFiles.length + materialsFiles.length + outputFiles.length,
+            timestamp: new Date().toISOString(),
+        };
+        console.log(`Total files found: ${result.totalFiles}`);
+        return {
+            content: [
+                {
+                    type: 'text',
+                    text: JSON.stringify(result),
+                },
+            ],
+        };
+    }
+    catch (error) {
+        return createErrorResponse(error, 'list-project-files');
+    }
+});
 server.registerTool('build-capcat-draft', {
     title: 'Build CapCut Draft',
     description: 'Read draft_content.json file, parse JSON and generate URIs for all assets in timeline tracks, then output the processed JSON string.',