Skip to main content
Glama
efikuta
by efikuta

analyze_comment_intents

Extract user intents and actionable insights from YouTube comments to understand audience engagement and feedback patterns.

Instructions

Analyze YouTube comments to extract user intents and actionable insights

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
videoIdYesYouTube video ID to analyze comments from
maxCommentsNoMaximum number of comments to analyze
intentCategoriesNoCustom intent categories to focus on (optional)

Implementation Reference

  • Entry point execute method of CommentIntentAnalyzer class that handles tool execution: parses args, checks cache, fetches comments, analyzes intents in batches using LLM, generates summary, caches result
    async execute(args: unknown): Promise<CommentIntentAnalysis> { const params = AnalyzeCommentIntentsSchema.parse(args); this.logger.info(`Analyzing comment intents for video: ${params.videoId}`); // Generate cache key const cacheKey = `comment_intents:${params.videoId}:${params.maxComments}:${JSON.stringify(params.intentCategories || [])}`; // Check cache first const cached = await this.cache.get<CommentIntentAnalysis>(cacheKey); if (cached) { this.logger.info(`Returning cached comment intent analysis for: ${params.videoId}`); return cached; } try { // Step 1: Get video details with comments const videoDetails = await this.youtubeClient.getVideoDetails({ videoId: params.videoId, includeTranscript: false, includeComments: true, maxComments: params.maxComments }); if (!videoDetails.comments || videoDetails.comments.length === 0) { throw new Error(`No comments found for video ${params.videoId}`); } // Step 2: Analyze comments in batches using LLM const intents = await this.analyzeCommentsInBatches( videoDetails.comments, params.intentCategories ); // Step 3: Generate summary and trends const analysis = await this.generateAnalysisSummary( params.videoId, intents, videoDetails.comments ); // Cache the result await this.cache.set(cacheKey, analysis, 3600); // 1 hour cache this.logger.info(`Comment intent analysis completed for ${params.videoId}: ${intents.length} intents identified`); return analysis; } catch (error) { this.logger.error(`Failed to analyze comment intents for ${params.videoId}:`, error); throw error; } }
  • Zod schema defining input parameters for the analyze_comment_intents tool
    export const AnalyzeCommentIntentsSchema = z.object({ videoId: z.string().describe('YouTube video ID'), maxComments: z.number().min(10).max(500).default(100).describe('Maximum comments to analyze'), intentCategories: z.array(z.string()).optional().describe('Custom intent categories to detect'), includeReplies: z.boolean().default(false).describe('Whether to include comment replies'), });
  • src/index.ts:415-441 (registration)
    Tool registration in listTools handler: name, description, and input schema
    name: 'analyze_comment_intents', description: 'Analyze YouTube comments to extract user intents and actionable insights', inputSchema: { type: 'object', properties: { videoId: { type: 'string', description: 'YouTube video ID to analyze comments from' }, maxComments: { type: 'number', minimum: 10, maximum: 200, default: 100, description: 'Maximum number of comments to analyze' }, intentCategories: { type: 'array', items: { type: 'string' }, description: 'Custom intent categories to focus on (optional)' } }, required: ['videoId'] } },
  • src/index.ts:584-586 (registration)
    Dispatch case in CallToolRequestSchema handler that routes to commentIntentTool.execute
    case 'analyze_comment_intents': result = await this.commentIntentTool.execute(args); break;
  • src/index.ts:179-179 (registration)
    Instantiation of CommentIntentAnalyzer class for use as the tool handler
    this.commentIntentTool = new CommentIntentAnalyzer(this.youtubeClient, this.cache, this.llmService, this.logger);

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/efikuta/youtube-knowledge-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server