Semantic Pen MCP Server

Instructions

Implementation Reference

• src/index.ts:535-579 (handler)

  The primary handler function that executes the 'get_article' tool logic. It makes an API request to fetch article details, processes the response (calculates word count, generates preview, formats status), and returns a structured MCP response with text content.

  private async getArticle(articleId: string) {
    const result = await this.makeRequest<ArticleDetail>(`/articles/${articleId}`);
    
    if (result.success && result.data) {
      const article = result.data;
      const title = article.extra_data?.targetArticleTopic || 'Untitled Article';
      const wordCount = article.output ? 
        Math.round(article.output.replace(/<[^>]*>/g, '').split(/\s+/).filter(word => word.length > 0).length) : 0;
      
      // Create a clean preview of the content (first 300 characters)
      const cleanContent = article.output ? 
        article.output
          .replace(/<[^>]*>/g, '') // Remove HTML tags
          .replace(/\n\s*\n/g, '\n') // Remove extra newlines
          .trim() 
        : 'Content not yet generated';
  
      const preview = cleanContent.length > 300 ? 
        cleanContent.substring(0, 300) + '...' : 
        cleanContent;
  
      const statusEmoji = article.status === 'finished' ? '✅' : 
                         article.status === 'processing' ? '🔄' : 
                         article.status === 'failed' ? '❌' : '⏳';
  
      return {
        content: [
          {
            type: "text",
            text: `📄 **${title}**\n\n${statusEmoji} **Status:** ${article.status} (${article.progress}%)\n**Article ID:** ${article.id}\n**Project:** ${article.project_name}\n**Created:** ${new Date(article.created_at).toLocaleDateString()}\n**Word Count:** ~${wordCount} words\n\n**Settings:**\n- Target Keyword: ${article.config?.targetKeyword || 'N/A'}\n- Language: ${article.config?.language || 'English'}\n- Type: ${article.config?.articleType || 'Article'}\n- Tone: ${article.config?.toneOfVoice || 'Professional'}\n- Target Words: ${article.config?.wordCount || 'N/A'}\n\n**Content Preview:**\n${preview}\n\n---\n💡 **Full HTML Content Available:** The complete article HTML is in the \`output\` field and can be used for publishing or further editing.`
          }
        ]
      };
    } else {
      return {
        content: [
          {
            type: "text",
            text: `❌ Failed to fetch article: ${result.error}`
          }
        ],
        isError: true
      };
    }
  }

• src/index.ts:272-281 (schema)

  The input schema for the 'get_article' tool, specifying that 'articleId' (string) is required.

  inputSchema: {
    type: "object",
    properties: {
      articleId: {
        type: "string",
        description: "The ID of the article to retrieve"
      }
    },
    required: ["articleId"]
  }

• src/index.ts:326-331 (registration)

  Registration and dispatch logic in the CallToolRequestHandler: validates input arguments and invokes the getArticle handler.

  case "get_article": {
    if (!args || typeof args !== 'object' || !('articleId' in args) || typeof args.articleId !== 'string') {
      throw new Error("articleId is required and must be a string");
    }
    return await this.getArticle(args.articleId);
  }

• src/index.ts:270-282 (schema)

  Full tool specification in ListToolsRequestHandler, including name, description, and input schema.

    name: "get_article",
    description: "Get a specific article by ID with full content",
    inputSchema: {
      type: "object",
      properties: {
        articleId: {
          type: "string",
          description: "The ID of the article to retrieve"
        }
      },
      required: ["articleId"]
    }
  }