@lutery/vision-mcp 1.0.0

package/README.md

@@ -0,0 +1,428 @@
+# Vision MCP
+
+MCP Server providing vision capabilities for LLMs via GLM-4.6V, SiliconFlow, and ModelScope. This server enables LLMs without native vision support or with expensive vision models to access cost-effective visual analysis capabilities.
+
+## Features
+
+- 🤖 **Multiple Model Support**: GLM-4.6V, SiliconFlow, and ModelScope vision models
+- 🖼️ **Flexible Image Input**: URL, base64 data URL, or local file paths
+- 📊 **Multiple Analysis Types**: Image description, UI analysis, object detection, OCR, and structured extraction
+- 🔧 **System Prompt Templates**: Built-in templates for common vision tasks
+- 📦 **Easy Deployment**: STDIO MCP Server, runs with npx
+- 🔒 **Secure**: Environment-based configuration, sensitive data masking in logs
+
+### Streaming Response Support
+
+Current adapters explicitly disable streaming responses (`stream: false`) and are designed for complete JSON responses. This ensures compatibility with both GLM-4.6V and SiliconFlow APIs.
+
+**Note**: Streaming-only providers are not currently supported. If a provider only supports streaming responses (Server-Sent Events/text/event-stream format), the adapter will fail as it expects a complete JSON response. To add support for streaming providers, a streaming response parser would need to be implemented.
+
+## Quick Start
+
+### Installation
+
+1. Clone or download this repository
+2. Install dependencies:
+
+```bash
+cd vision_mcp
+npm install
+```
+
+### Configuration
+
+Create a `.env` file in the project root:
+
+#### Option 1: GLM-4.6V
+
+```bash
+VISION_MODEL_TYPE=glm-4.6v
+VISION_MODEL_NAME=glm-4.6v
+VISION_API_BASE_URL=https://open.bigmodel.cn/api/paas/v4
+VISION_API_KEY=your-glm-api-key
+```
+
+#### Option 2: SiliconFlow
+
+```bash
+VISION_MODEL_TYPE=siliconflow
+VISION_MODEL_NAME=Qwen/Qwen2-VL-72B-Instruct
+VISION_API_BASE_URL=https://api.siliconflow.cn/v1
+VISION_API_KEY=your-siliconflow-api-key
+```
+
+#### Option 3: ModelScope API-Inference
+
+```bash
+VISION_MODEL_TYPE=modelscope
+VISION_MODEL_NAME=ZhipuAI/GLM-4.6V
+VISION_API_BASE_URL=https://api-inference.modelscope.cn/v1
+VISION_API_KEY=your-modelscope-token
+```
+
+**Note**: ModelScope requires:
+- Real-name authentication on your ModelScope account
+- Aliyun account binding
+- API usage limits apply (see [API Limits](https://www.modelscope.cn/docs/model-service/API-Inference/limits))
+
+### Build
+
+```bash
+npm run build
+```
+
+### Run (local)
+
+```bash
+node dist/index.js
+```
+
+If successful, you'll see: `Vision MCP Server is running on stdio` in stderr.
+
+### Run (npx)
+
+```bash
+# Local package (requires build first)
+npx .
+
+# Published package
+npx -y @lutery/vision-mcp
+```
+
+## MCP Client Configuration
+
+### Claude Desktop
+
+Add to your Claude Desktop configuration:
+
+```json
+{
+  "mcpServers": {
+    "vision-mcp": {
+      "command": "npx",
+      "args": ["-y", "@lutery/vision-mcp"],
+      "env": {
+        "VISION_MODEL_TYPE": "glm-4.6v",
+        "VISION_MODEL_NAME": "glm-4.6v",
+        "VISION_API_BASE_URL": "https://open.bigmodel.cn/api/paas/v4",
+        "VISION_API_KEY": "your-api-key"
+      }
+    }
+  }
+}
+```
+
+Or with a local installation:
+
+```json
+{
+  "mcpServers": {
+    "vision-mcp": {
+      "command": "node",
+      "args": ["/path/to/vision_mcp/dist/index.js"],
+      "env": {
+        "VISION_MODEL_TYPE": "glm-4.6v",
+        "VISION_MODEL_NAME": "glm-4.6v",
+        "VISION_API_BASE_URL": "https://open.bigmodel.cn/api/paas/v4",
+        "VISION_API_KEY": "your-api-key"
+      }
+    }
+  }
+}
+```
+
+### Cursor/Codex CLI
+
+Similar configuration for other MCP-compatible clients.
+
+## Using the Tools
+
+### 1. Analyze Image
+
+Main tool for image analysis:
+
+```javascript
+// Tool: analyze_image
+// Parameters:
+{
+  "image": "https://example.com/image.jpg",        // Image URL, base64, or local path
+  "prompt": "Describe this UI design in detail",   // Analysis prompt
+  "output_format": "text",                          // Optional: "text" or "json"
+  "template": "ui-analysis"                         // Optional: see templates below
+}
+```
+
+#### Example Prompts
+
+**UI Analysis:**
+```json
+{
+  "image": "./screenshot.png",
+  "prompt": "Analyze this UI design and extract all UI components with their positions and styles",
+  "template": "ui-analysis"
+}
+```
+
+**Object Detection:**
+```json
+{
+  "image": "https://example.com/photo.jpg",
+  "prompt": "Detect all objects and provide their coordinates",
+  "template": "object-detection"
+}
+```
+
+**OCR:**
+```json
+{
+  "image": "data:image/png;base64,iVBORw0KGgo...",
+  "prompt": "Extract all text from this image",
+  "template": "ocr"
+}
+```
+
+**Structured Extraction:**
+```json
+{
+  "image": "./form.jpg",
+  "prompt": "Extract all form fields and values as JSON",
+  "output_format": "json"
+}
+```
+
+### 2. List Templates
+
+List available system prompt templates:
+
+```javascript
+// Tool: list_templates
+// Parameters: none
+```
+
+Available templates:
+- `general-description` - General image description
+- `ui-analysis` - UI prototype and interface analysis
+- `object-detection` - Object detection and localization
+- `ocr` - Text extraction (OCR)
+- `structured-extraction` - Structured data extraction
+
+### 3. Get Config
+
+Get current model configuration:
+
+```javascript
+// Tool: get_config
+// Parameters: none
+```
+
+## Image Input Formats
+
+### 1. URL
+
+```
+https://example.com/image.jpg
+```
+
+### 2. Base64 Data URL
+
+```
+data:image/jpeg;base64,/9j/4AAQSkZJRgABAQAAAQABAAD...
+```
+
+### 3. Local File Path
+
+```
+/path/to/image.png
+./relative/path/image.jpg
+```
+
+Note: Local paths only work if the MCP server has access to the filesystem.
+Note: URL validation is strict by default (see `VISION_STRICT_URL_VALIDATION`).
+
+## Environment Variables
+
+| Variable | Description | Default | Required |
+|----------|-------------|---------|----------|
+| `VISION_MODEL_TYPE` | Model type: `glm` (alias for `glm-4.6v`), `glm-4.6v`, `siliconflow`, or `modelscope` | - | Yes |
+| `VISION_MODEL_NAME` | Model name for the API | See defaults below | Yes |
+| `VISION_API_BASE_URL` | API base URL (must be base path, no `/chat/completions`) | See defaults below | Yes |
+| `VISION_API_KEY` | API key for authentication | - | Yes |
+| `VISION_API_TIMEOUT` | Request timeout in milliseconds | 60000 | No |
+| `VISION_MAX_RETRIES` | Maximum retry attempts | 2 | No |
+| `VISION_STRICT_URL_VALIDATION` | Enforce strict image URL validation | `true` | No |
+| `LOG_LEVEL` | Log level: `debug`, `info`, `warn`, `error` | `info` | No |
+
+**Notes**:
+- `VISION_STRICT_URL_VALIDATION` defaults to `true`, enforcing strict validation that URLs must end with supported image extensions (`.jpg`, `.jpeg`, `.png`, `.webp`). Set to `false` to allow non-image URLs with a warning only.
+- For GLM-4.6V provider, both `glm` and `glm-4.6v` values work for `VISION_MODEL_TYPE`. `glm` is provided as a convenient alias.
+
+### Model Defaults
+
+**GLM-4.6V:**
+```bash
+VISION_MODEL_NAME=glm-4.6v
+VISION_API_BASE_URL=https://open.bigmodel.cn/api/paas/v4
+```
+
+**SiliconFlow:**
+```bash
+VISION_MODEL_NAME=Qwen/Qwen2-VL-72B-Instruct
+VISION_API_BASE_URL=https://api.siliconflow.cn/v1
+```
+
+## API Keys
+
+### GLM-4.6V
+
+Get your API key from: [智谱 AI 开放平台](https://open.bigmodel.cn/)
+
+Format: `xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx.xxxxxxxxxxxxxxxxxxxx`
+
+### SiliconFlow
+
+Get your API key from: [SiliconFlow](https://cloud.siliconflow.cn/)
+
+Format: `sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx`
+
+## MCP Protocol Note
+
+**IMPORTANT**: This is a STDIO-based MCP Server. According to MCP protocol:
+
+- **DO NOT** use `console.log()` or write to stdout
+- **USE ONLY** `console.error()` for logging (stderr)
+- stdout is reserved for JSON-RPC communication
+
+The server handles this automatically. If you fork this project, ensure you follow this rule.
+
+## Development
+
+### Project Structure
+
+```
+vision_mcp/
+├── src/
+│   ├── index.ts              # MCP Server entry point
+│   ├── config/
+│   │   └── model-config.ts   # Configuration management
+│   ├── tools/
+│   │   └── vision-tool.ts    # Vision analysis tool
+│   ├── adapters/
+│   │   ├── base-adapter.ts   # Base adapter class
+│   │   ├── glm-adapter.ts    # GLM-4.6V adapter
+│   │   └── siliconflow-adapter.ts  # SiliconFlow adapter
+│   ├── prompts/
+│   │   └── system.ts         # System prompt templates
+│   └── utils/
+│       ├── errors.ts         # Error handling
+│       ├── logger.ts         # Logging utilities
+│       └── image-input.ts    # Image input normalization
+├── package.json
+├── tsconfig.json
+└── README.md
+```
+
+### Building
+
+```bash
+# Install dependencies
+npm install
+
+# Build TypeScript
+npm run build
+
+# Run tests
+npm test
+```
+
+### Testing Notes
+
+- `npm test` uses `VISION_API_KEY` (default) or provider-specific keys in the test script:
+  - `SILICONFLOW_API_KEY`
+  - `GLM_API_KEY`
+- If no API key is set, the tests will exit with a clear error message.
+
+## Troubleshooting
+
+### 1. "Failed to load model configuration"
+
+- Check all required environment variables are set
+- Verify `VISION_MODEL_TYPE` is either `glm-4.6v` or `siliconflow`
+
+### 2. "API Key not found"
+
+- Set `VISION_API_KEY` in your environment
+- Verify the API key format matches the model requirements
+
+### 3. "Connection timeout"
+
+- Increase `VISION_API_TIMEOUT` value
+- Check network connectivity to the API endpoint
+- Verify API endpoint URL is correct
+
+### 4. "Invalid image URL"
+
+- Ensure URL is publicly accessible
+- Check URL format (http:// or https://)
+- Verify image format is supported
+
+### 5. "Permission denied reading file"
+
+- MCP server needs filesystem access for local files
+- Use absolute paths or ensure relative paths are accessible
+- Check file permissions
+
+### 6. "Invalid API endpoint" or "404 Not Found"
+
+- Ensure `VISION_API_BASE_URL` is the base path only, without `/chat/completions`
+- Correct: `https://api.siliconflow.cn/v1`
+- Incorrect: `https://api.siliconflow.cn/v1/chat/completions`
+- Check the error details for the full request URL to diagnose endpoint issues
+
+## Security Notes
+
+- API keys are loaded from environment variables, never hardcoded
+- API keys are masked in logs
+- Images are not persisted by default
+- MCP server should run in trusted environments only (no built-in auth)
+- **Thinking/Reasoning Content Filtering**: Model thinking/reasoning content is automatically filtered from responses to prevent exposing internal reasoning to MCP clients. This filtering is unconditional and applied to all supported models regardless of configuration.
+
+## Security Best Practices
+
+⚠️ **IMPORTANT**: Never commit API keys or credentials to the repository!
+
+- **Use environment variables** for sensitive data (`.env` file)
+- **Keep local test credentials** in `.gitignore`'d files (e.g., `test_key.local.md`)
+- **Rotate keys immediately** if accidentally exposed or committed
+- **See** `doc/test_key.example.md` for test setup template
+- **Never** copy real API keys into documentation, code comments, or issue trackers
+
+**Key Protection Checklist**:
+- [ ] `.env` is in `.gitignore`
+- [ ] `.env.local` is in `.gitignore`
+- [ ] No real keys in `test_key.md` (use `test_key.example.md` instead)
+- [ ] No keys in documentation or comments
+- [ ] Review git history for accidental key commits (`git log --all --full-history -S --source --all -- "*secret*" "*key*" "*password*" "test_key.md"`)
+
+## License
+
+MIT
+
+## Contributing
+
+1. Fork the repository
+2. Create a feature branch
+3. Make your changes
+4. Add tests
+5. Submit a pull request
+
+## Support
+
+For issues and questions:
+- Open an issue on the repository
+- Check model documentation:
+  - [GLM-4.6V Docs](https://docs.bigmodel.cn/)
+  - [SiliconFlow Docs](https://docs.siliconflow.cn/)
+
+
+## TODO
+- [ ] 适配modelscope的视觉模型接口请求：https://www.modelscope.cn/docs/model-service/API-Inference/intro

package/dist/adapters/base-adapter.d.ts

@@ -0,0 +1,69 @@
+/**
+ * Base Vision Model Adapter
+ *
+ * @description 定义模型适配器的统一接口和抽象基类
+ */
+import { ModelConfig } from '../config/model-config.js';
+/**
+ * 模型响应接口
+ */
+export interface VisionModelResponse {
+    content: string;
+    usage?: {
+        promptTokens?: number;
+        completionTokens?: number;
+        totalTokens?: number;
+    };
+    model?: string;
+}
+/**
+ * 模型适配器接口
+ */
+export interface VisionModelAdapter {
+    config: ModelConfig;
+    /**
+     * 分析图片
+     * @param imageData - 图片数据（URL 或 base64）
+     * @param prompt - 提示词
+     * @returns 模型响应
+     */
+    analyze(imageData: string, prompt: string): Promise<string>;
+    /**
+     * 分析图片（带完整响应）
+     * @param imageData - 图片数据（URL 或 base64）
+     * @param prompt - 提示词
+     * @returns 完整响应
+     */
+    analyzeWithResponse(imageData: string, prompt: string): Promise<VisionModelResponse>;
+}
+/**
+ * 抽象基类实现通用功能
+ */
+export declare abstract class BaseVisionModelAdapter implements VisionModelAdapter {
+    config: ModelConfig;
+    constructor(config: ModelConfig);
+    abstract analyze(imageData: string, prompt: string): Promise<string>;
+    abstract analyzeWithResponse(imageData: string, prompt: string): Promise<VisionModelResponse>;
+    /**
+     * 不可重试的 HTTP 状态码
+     * 这些错误表示客户端配置问题，重试不会改变结果
+     */
+    private readonly NON_RETRYABLE_STATUS_CODES;
+    /**
+     * 带重试和超时控制的请求包装器
+     *
+     * @note 当前实现强制使用非流式响应（stream: false）。
+     *       如果提供商只支持流式响应，需要添加流式解析器。
+     */
+    protected withRetry<T>(operation: (signal: AbortSignal) => Promise<T>, config?: {
+        maxRetries?: number;
+        timeout?: number;
+    }): Promise<T>;
+    /**
+     * 解析模型响应
+     * @param response - 原始响应
+     * @param modelType - 模型类型（必需，用于过滤 thinking content）
+     */
+    protected parseResponse(response: unknown, modelType: string): VisionModelResponse;
+}
+//# sourceMappingURL=base-adapter.d.ts.map

package/dist/adapters/base-adapter.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"base-adapter.d.ts","sourceRoot":"","sources":["../../src/adapters/base-adapter.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAE,WAAW,EAAE,MAAM,2BAA2B,CAAC;AAKxD;;GAEG;AACH,MAAM,WAAW,mBAAmB;IAClC,OAAO,EAAE,MAAM,CAAC;IAChB,KAAK,CAAC,EAAE;QACN,YAAY,CAAC,EAAE,MAAM,CAAC;QACtB,gBAAgB,CAAC,EAAE,MAAM,CAAC;QAC1B,WAAW,CAAC,EAAE,MAAM,CAAC;KACtB,CAAC;IACF,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAED;;GAEG;AACH,MAAM,WAAW,kBAAkB;IACjC,MAAM,EAAE,WAAW,CAAC;IAEpB;;;;;OAKG;IACH,OAAO,CAAC,SAAS,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;IAE5D;;;;;OAKG;IACH,mBAAmB,CAAC,SAAS,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,mBAAmB,CAAC,CAAC;CACtF;AAED;;GAEG;AACH,8BAAsB,sBAAuB,YAAW,kBAAkB;IACrD,MAAM,EAAE,WAAW;gBAAnB,MAAM,EAAE,WAAW;IAEtC,QAAQ,CAAC,OAAO,CAAC,SAAS,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC;IACpE,QAAQ,CAAC,mBAAmB,CAAC,SAAS,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,mBAAmB,CAAC;IAE7F;;;OAGG;IACH,OAAO,CAAC,QAAQ,CAAC,0BAA0B,CAAiC;IAE5E;;;;;OAKG;cACa,SAAS,CAAC,CAAC,EACzB,SAAS,EAAE,CAAC,MAAM,EAAE,WAAW,KAAK,OAAO,CAAC,CAAC,CAAC,EAC9C,MAAM,GAAE;QAAE,UAAU,CAAC,EAAE,MAAM,CAAC;QAAC,OAAO,CAAC,EAAE,MAAM,CAAA;KAAO,GACrD,OAAO,CAAC,CAAC,CAAC;IA+Eb;;;;OAIG;IACH,SAAS,CAAC,aAAa,CAAC,QAAQ,EAAE,OAAO,EAAE,SAAS,EAAE,MAAM,GAAG,mBAAmB;CAuDnF"}

package/dist/adapters/base-adapter.js

@@ -0,0 +1,143 @@
+/**
+ * Base Vision Model Adapter
+ *
+ * @description 定义模型适配器的统一接口和抽象基类
+ */
+import { ModelAPIError, TimeoutError, VisionMCPError } from '../utils/errors.js';
+import { logger } from '../utils/logger.js';
+import { filterThinkingContent } from '../utils/thinking-filter.js';
+/**
+ * 抽象基类实现通用功能
+ */
+export class BaseVisionModelAdapter {
+    config;
+    constructor(config) {
+        this.config = config;
+    }
+    /**
+     * 不可重试的 HTTP 状态码
+     * 这些错误表示客户端配置问题，重试不会改变结果
+     */
+    NON_RETRYABLE_STATUS_CODES = new Set([400, 401, 403, 404]);
+    /**
+     * 带重试和超时控制的请求包装器
+     *
+     * @note 当前实现强制使用非流式响应（stream: false）。
+     *       如果提供商只支持流式响应，需要添加流式解析器。
+     */
+    async withRetry(operation, config = {}) {
+        const { maxRetries = this.config.maxRetries || 2, timeout = this.config.timeout || 60000 } = config;
+        let lastError;
+        for (let attempt = 0; attempt <= maxRetries; attempt++) {
+            try {
+                logger.debug(`Attempt ${attempt + 1}/${maxRetries + 1}`);
+                // 使用 AbortController 实现超时
+                const controller = new AbortController();
+                const timeoutId = setTimeout(() => controller.abort(), timeout);
+                try {
+                    const result = await operation(controller.signal);
+                    clearTimeout(timeoutId);
+                    return result;
+                }
+                catch (error) {
+                    clearTimeout(timeoutId);
+                    throw error;
+                }
+            }
+            catch (error) {
+                lastError = error;
+                // 如果是 AbortError，转换为 TimeoutError
+                if (error instanceof Error && error.name === 'AbortError') {
+                    throw new TimeoutError(`Request timed out after ${timeout}ms`);
+                }
+                // 检查是否为不可重试错误（400/401/403/404）
+                if (error instanceof ModelAPIError) {
+                    const status = error.details?.status;
+                    if (status && this.NON_RETRYABLE_STATUS_CODES.has(status)) {
+                        logger.error('Non-retryable error, failing immediately', {
+                            status,
+                            attempt: attempt + 1
+                        });
+                        throw error; // 直接抛出，不重试
+                    }
+                }
+                // 最后一次尝试失败，抛出错误
+                if (attempt === maxRetries) {
+                    break;
+                }
+                // 计算退避时间（指数退避）
+                const backoffTime = Math.min(1000 * Math.pow(2, attempt), 5000);
+                logger.warn(`Attempt ${attempt + 1} failed, retrying in ${backoffTime}ms`, {
+                    error: lastError.message
+                });
+                // 等待后退时间
+                await new Promise(resolve => setTimeout(resolve, backoffTime));
+            }
+        }
+        // 保留最后失败的完整错误详情（如果可用）
+        let errorDetails;
+        if (lastError !== undefined && lastError instanceof VisionMCPError && lastError.details) {
+            // 如果是 VisionMCPError，保留所有 details
+            errorDetails = {
+                ...lastError.details,
+                lastError: lastError.message
+            };
+        }
+        else {
+            // 否则只保留消息
+            errorDetails = { lastError: lastError?.message };
+        }
+        throw new ModelAPIError(`Failed after ${maxRetries + 1} attempts`, errorDetails);
+    }
+    /**
+     * 解析模型响应
+     * @param response - 原始响应
+     * @param modelType - 模型类型（必需，用于过滤 thinking content）
+     */
+    parseResponse(response, modelType) {
+        // modelType is required
+        if (!modelType) {
+            throw new Error('modelType is required for parseResponse');
+        }
+        try {
+            // @ts-ignore - 检查响应结构
+            const content = response?.choices?.[0]?.message?.content;
+            if (!content || typeof content !== 'string') {
+                throw new ModelAPIError('Invalid response format: missing or invalid content', { response });
+            }
+            // 过滤 thinking/reasoning content（无条件执行）
+            let filteredContent;
+            try {
+                filteredContent = filterThinkingContent(response, modelType);
+                // 如果有 content 被过滤，记录日志
+                if (filteredContent.length < content.length) {
+                    logger.debug('Filtered thinking content from response', {
+                        modelType,
+                        originalLength: content.length,
+                        filteredLength: filteredContent.length,
+                        reduction: content.length - filteredContent.length
+                    });
+                }
+            }
+            catch (error) {
+                logger.warn('Failed to filter thinking content, returning raw content', {
+                    modelType,
+                    error: error instanceof Error ? error.message : error
+                });
+                filteredContent = content;
+            }
+            return {
+                content: filteredContent,
+                // @ts-ignore
+                usage: response?.usage,
+                // @ts-ignore
+                model: response?.model
+            };
+        }
+        catch (error) {
+            logger.error('Failed to parse model response', error);
+            throw new ModelAPIError('Failed to parse model response', { response }, error);
+        }
+    }
+}
+//# sourceMappingURL=base-adapter.js.map

package/dist/adapters/base-adapter.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"base-adapter.js","sourceRoot":"","sources":["../../src/adapters/base-adapter.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAGH,OAAO,EAAE,aAAa,EAAE,YAAY,EAAE,cAAc,EAAE,MAAM,oBAAoB,CAAC;AACjF,OAAO,EAAE,MAAM,EAAE,MAAM,oBAAoB,CAAC;AAC5C,OAAO,EAAE,qBAAqB,EAAE,MAAM,6BAA6B,CAAC;AAsCpE;;GAEG;AACH,MAAM,OAAgB,sBAAsB;IACvB;IAAnB,YAAmB,MAAmB;QAAnB,WAAM,GAAN,MAAM,CAAa;IAAG,CAAC;IAK1C;;;OAGG;IACc,0BAA0B,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;IAE5E;;;;;OAKG;IACO,KAAK,CAAC,SAAS,CACvB,SAA8C,EAC9C,SAAoD,EAAE;QAEtD,MAAM,EACJ,UAAU,GAAG,IAAI,CAAC,MAAM,CAAC,UAAU,IAAI,CAAC,EACxC,OAAO,GAAG,IAAI,CAAC,MAAM,CAAC,OAAO,IAAI,KAAK,EACvC,GAAG,MAAM,CAAC;QAEX,IAAI,SAA4B,CAAC;QAEjC,KAAK,IAAI,OAAO,GAAG,CAAC,EAAE,OAAO,IAAI,UAAU,EAAE,OAAO,EAAE,EAAE,CAAC;YACvD,IAAI,CAAC;gBACH,MAAM,CAAC,KAAK,CAAC,WAAW,OAAO,GAAG,CAAC,IAAI,UAAU,GAAG,CAAC,EAAE,CAAC,CAAC;gBAEzD,0BAA0B;gBAC1B,MAAM,UAAU,GAAG,IAAI,eAAe,EAAE,CAAC;gBACzC,MAAM,SAAS,GAAG,UAAU,CAAC,GAAG,EAAE,CAAC,UAAU,CAAC,KAAK,EAAE,EAAE,OAAO,CAAC,CAAC;gBAEhE,IAAI,CAAC;oBACH,MAAM,MAAM,GAAG,MAAM,SAAS,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC;oBAClD,YAAY,CAAC,SAAS,CAAC,CAAC;oBACxB,OAAO,MAAM,CAAC;gBAChB,CAAC;gBAAC,OAAO,KAAK,EAAE,CAAC;oBACf,YAAY,CAAC,SAAS,CAAC,CAAC;oBACxB,MAAM,KAAK,CAAC;gBACd,CAAC;YACH,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,SAAS,GAAG,KAAc,CAAC;gBAE3B,kCAAkC;gBAClC,IAAI,KAAK,YAAY,KAAK,IAAI,KAAK,CAAC,IAAI,KAAK,YAAY,EAAE,CAAC;oBAC1D,MAAM,IAAI,YAAY,CAAC,2BAA2B,OAAO,IAAI,CAAC,CAAC;gBACjE,CAAC;gBAED,+BAA+B;gBAC/B,IAAI,KAAK,YAAY,aAAa,EAAE,CAAC;oBACnC,MAAM,MAAM,GAAI,KAAK,CAAC,OAAe,EAAE,MAAM,CAAC;oBAC9C,IAAI,MAAM,IAAI,IAAI,CAAC,0BAA0B,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC;wBAC1D,MAAM,CAAC,KAAK,CAAC,0CAA0C,EAAE;4BACvD,MAAM;4BACN,OAAO,EAAE,OAAO,GAAG,CAAC;yBACrB,CAAC,CAAC;wBACH,MAAM,KAAK,CAAC,CAAC,WAAW;oBAC1B,CAAC;gBACH,CAAC;gBAED,gBAAgB;gBAChB,IAAI,OAAO,KAAK,UAAU,EAAE,CAAC;oBAC3B,MAAM;gBACR,CAAC;gBAED,eAAe;gBACf,MAAM,WAAW,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,OAAO,CAAC,EAAE,IAAI,CAAC,CAAC;gBAChE,MAAM,CAAC,IAAI,CAAC,WAAW,OAAO,GAAG,CAAC,wBAAwB,WAAW,IAAI,EAAE;oBACzE,KAAK,EAAE,SAAS,CAAC,OAAO;iBACzB,CAAC,CAAC;gBAEH,SAAS;gBACT,MAAM,IAAI,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,WAAW,CAAC,CAAC,CAAC;YACjE,CAAC;QACH,CAAC;QAED,sBAAsB;QACtB,IAAI,YAAiC,CAAC;QACtC,IAAI,SAAS,KAAK,SAAS,IAAI,SAAS,YAAY,cAAc,IAAI,SAAS,CAAC,OAAO,EAAE,CAAC;YACxF,kCAAkC;YAClC,YAAY,GAAG;gBACb,GAAG,SAAS,CAAC,OAAO;gBACpB,SAAS,EAAE,SAAS,CAAC,OAAO;aAC7B,CAAC;QACJ,CAAC;aAAM,CAAC;YACN,UAAU;YACV,YAAY,GAAG,EAAE,SAAS,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC;QACnD,CAAC;QAED,MAAM,IAAI,aAAa,CACrB,gBAAgB,UAAU,GAAG,CAAC,WAAW,EACzC,YAAY,CACb,CAAC;IACJ,CAAC;IAED;;;;OAIG;IACO,aAAa,CAAC,QAAiB,EAAE,SAAiB;QAC1D,wBAAwB;QACxB,IAAI,CAAC,SAAS,EAAE,CAAC;YACf,MAAM,IAAI,KAAK,CAAC,yCAAyC,CAAC,CAAC;QAC7D,CAAC;QAED,IAAI,CAAC;YACH,sBAAsB;YACtB,MAAM,OAAO,GAAG,QAAQ,EAAE,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,OAAO,CAAC;YAEzD,IAAI,CAAC,OAAO,IAAI,OAAO,OAAO,KAAK,QAAQ,EAAE,CAAC;gBAC5C,MAAM,IAAI,aAAa,CACrB,qDAAqD,EACrD,EAAE,QAAQ,EAAE,CACb,CAAC;YACJ,CAAC;YAED,uCAAuC;YACvC,IAAI,eAAuB,CAAC;YAC5B,IAAI,CAAC;gBACH,eAAe,GAAG,qBAAqB,CAAC,QAAQ,EAAE,SAAS,CAAC,CAAC;gBAE7D,uBAAuB;gBACvB,IAAI,eAAe,CAAC,MAAM,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC;oBAC5C,MAAM,CAAC,KAAK,CAAC,yCAAyC,EAAE;wBACtD,SAAS;wBACT,cAAc,EAAE,OAAO,CAAC,MAAM;wBAC9B,cAAc,EAAE,eAAe,CAAC,MAAM;wBACtC,SAAS,EAAE,OAAO,CAAC,MAAM,GAAG,eAAe,CAAC,MAAM;qBACnD,CAAC,CAAC;gBACL,CAAC;YACH,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,MAAM,CAAC,IAAI,CAAC,0DAA0D,EAAE;oBACtE,SAAS;oBACT,KAAK,EAAE,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,KAAK;iBACtD,CAAC,CAAC;gBACH,eAAe,GAAG,OAAO,CAAC;YAC5B,CAAC;YAED,OAAO;gBACL,OAAO,EAAE,eAAe;gBACxB,aAAa;gBACb,KAAK,EAAE,QAAQ,EAAE,KAAK;gBACtB,aAAa;gBACb,KAAK,EAAE,QAAQ,EAAE,KAAK;aACvB,CAAC;QACJ,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,MAAM,CAAC,KAAK,CAAC,gCAAgC,EAAE,KAAK,CAAC,CAAC;YACtD,MAAM,IAAI,aAAa,CACrB,gCAAgC,EAChC,EAAE,QAAQ,EAAE,EACZ,KAAK,CACN,CAAC;QACJ,CAAC;IACH,CAAC;CACF"}

package/dist/adapters/claude-adapter.d.ts

@@ -0,0 +1,38 @@
+/**
+ * Claude (Anthropic Messages API) Adapter
+ *
+ * @description Claude Messages API 适配器实现，支持 Claude 多模态视觉模型
+ * @see https://docs.anthropic.com/claude/reference/messages-post
+ */
+import { BaseVisionModelAdapter, VisionModelResponse } from './base-adapter.js';
+import { ModelConfig } from '../config/model-config.js';
+export interface ClaudeAdapterOptions {
+    maxTokens?: number;
+    apiVersion?: string;
+}
+export declare class ClaudeAdapter extends BaseVisionModelAdapter {
+    private options;
+    constructor(config: ModelConfig, options?: ClaudeAdapterOptions);
+    analyze(imageData: string, prompt: string): Promise<string>;
+    analyzeWithResponse(imageData: string, prompt: string): Promise<VisionModelResponse>;
+    private callClaudeAPI;
+    /**
+     * 构建图片 content block
+     * 支持 URL 和 base64 data URL 两种格式
+     */
+    private buildImageBlock;
+    /**
+     * 处理错误响应
+     */
+    private handleErrorResponse;
+    /**
+     * 解析响应数据
+     */
+    private parseResponseData;
+    /**
+     * 归一化 Claude 响应为 VisionModelResponse
+     * Claude 响应格式：{ content: [{type: "text", text: "..."}], usage: {...} }
+     */
+    private normalizeResponse;
+}
+//# sourceMappingURL=claude-adapter.d.ts.map

package/dist/adapters/claude-adapter.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"claude-adapter.d.ts","sourceRoot":"","sources":["../../src/adapters/claude-adapter.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,EAAE,sBAAsB,EAAE,mBAAmB,EAAE,MAAM,mBAAmB,CAAC;AAChF,OAAO,EAAE,WAAW,EAAE,MAAM,2BAA2B,CAAC;AAKxD,MAAM,WAAW,oBAAoB;IACnC,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;AAED,qBAAa,aAAc,SAAQ,sBAAsB;IACvD,OAAO,CAAC,OAAO,CAAiC;gBAEpC,MAAM,EAAE,WAAW,EAAE,OAAO,GAAE,oBAAyB;IAyB7D,OAAO,CAAC,SAAS,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC;IAyB3D,mBAAmB,CAAC,SAAS,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,mBAAmB,CAAC;YA0B5E,aAAa;IAgE3B;;;OAGG;IACH,OAAO,CAAC,eAAe;IAwBvB;;OAEG;YACW,mBAAmB;IAwCjC;;OAEG;YACW,iBAAiB;IAiC/B;;;OAGG;IACH,OAAO,CAAC,iBAAiB;CA+B1B"}