content-optimizer-mcp

Instructions

Implementation Reference

• mcp-server/src/index.ts:372-447 (handler)

  The `get_content_recommendations` tool is registered and implemented within `mcp-server/src/index.ts`. It takes `content` and `keyword` as inputs, uses `scoreContent`, `analyzeSERP`, and `countWords` to evaluate the content, then generates and prioritizes a list of recommendations based on SERP benchmarks and content scores.

  server.tool(
    "get_content_recommendations",
    "Get a prioritized list of specific improvements to optimize content for a target keyword.",
    {
      content: z.string().describe("The content to analyze"),
      keyword: z.string().describe("The target keyword"),
    },
    async ({ content, keyword }) => {
      const { overallScore, categories, recommendations } = scoreContent(content, keyword);
      const serp = analyzeSERP(keyword);
      const words = countWords(content);
  
      const prioritized: { priority: "high" | "medium" | "low"; recommendation: string }[] = [];
  
      // High priority: low-scoring categories
      for (const [name, cat] of Object.entries(categories)) {
        const ratio = cat.score / cat.maxScore;
        if (ratio < 0.4) {
          prioritized.push({ priority: "high", recommendation: `[${name}] ${cat.details}` });
        }
      }
  
      // Word count gap
      if (words < serp.averages.recommendedWordCount * 0.7) {
        prioritized.push({
          priority: "high",
          recommendation: `Content is ${words} words. Top-ranking pages average ${serp.averages.wordCount} words. Aim for ${serp.averages.recommendedWordCount}+.`,
        });
      }
  
      // Add scored recommendations
      for (const rec of recommendations) {
        const isHigh = rec.toLowerCase().includes("keyword") && rec.toLowerCase().includes("first paragraph");
        prioritized.push({ priority: isHigh ? "high" : "medium", recommendation: rec });
      }
  
      // Missing topics
      const topics = getKeywordTopics(keyword);
      const contentLower = content.toLowerCase();
      const missing = topics.filter((t) => !contentLower.includes(t.toLowerCase()));
      if (missing.length > topics.length * 0.5) {
        prioritized.push({
          priority: "high",
          recommendation: `Missing ${missing.length}/${topics.length} key topics. Add sections on: ${missing.slice(0, 5).join(", ")}.`,
        });
      }
  
      // Deduplicate
      const seen = new Set<string>();
      const unique = prioritized.filter((r) => {
        if (seen.has(r.recommendation)) return false;
        seen.add(r.recommendation);
        return true;
      });
  
      // Sort by priority
      const order = { high: 0, medium: 1, low: 2 };
      unique.sort((a, b) => order[a.priority] - order[b.priority]);
  
      return {
        content: [
          {
            type: "text" as const,
            text: JSON.stringify({
              overallScore,
              targetKeyword: keyword,
              recommendations: unique,
              serpBenchmark: {
                avgWordCount: serp.averages.wordCount,
                recommendedWordCount: serp.averages.recommendedWordCount,
                avgHeadings: serp.averages.headingCount,
              },
            }, null, 2),
          },
        ],
      };