LinkedIn MCP Server

Instructions

Implementation Reference

• src/index.ts:842-870 (handler)

  The main handler function that implements the search_jobs tool. Builds query parameters from input args, calls the HarvestAPI /job-search endpoint, cleans the job search results using DataCleaners.cleanJobSearchResult, limits results, and returns a formatted TOON-encoded response with optional saving and pagination info.
  private async searchJobs(args: Record<string, any>): Promise<CallToolResult> { const params: Record<string, any> = {}; if (args.search) params.search = args.search; if (args.companyId) params.companyId = args.companyId; if (args.location) params.location = args.location; if (args.geoId) params.geoId = args.geoId; if (args.sortBy) params.sortBy = args.sortBy; if (args.workplaceType) params.workplaceType = args.workplaceType; if (args.employmentType) params.employmentType = args.employmentType; if (args.salary) params.salary = args.salary; if (args.postedLimit) params.postedLimit = args.postedLimit; if (args.experienceLevel) params.experienceLevel = args.experienceLevel; if (args.industryId) params.industryId = args.industryId; if (args.functionId) params.functionId = args.functionId; if (args.under10Applicants) params.under10Applicants = args.under10Applicants; if (args.easyApply !== undefined) params.easyApply = args.easyApply; if (args.page) params.page = args.page; const data = await this.makeRequest('/job-search', params); const maxItems = args.max_items || 10; const cleaned = (data.elements || []).slice(0, maxItems).map(DataCleaners.cleanJobSearchResult); return this.formatResponse(cleaned, { saveDir: args.save_dir, toolName: 'search_jobs', pagination: data.pagination, }); }
• src/index.ts:394-416 (schema)

  Input schema definition for the search_jobs tool, specifying all available parameters, their types, descriptions, enums, and defaults for MCP validation.
  inputSchema: { type: 'object', properties: { search: { type: 'string', description: 'Search jobs by title' }, companyId: { type: 'string', description: 'Filter by company ID' }, location: { type: 'string', description: 'Filter by location' }, geoId: { type: 'string', description: 'Filter by LinkedIn Geo ID' }, sortBy: { type: 'string', description: 'Sort by: relevance or date', enum: ['relevance', 'date'] }, workplaceType: { type: 'string', description: 'Filter: office, hybrid, remote', enum: ['office', 'hybrid', 'remote'] }, employmentType: { type: 'string', description: 'Filter: full-time, part-time, contract, temporary, volunteer, internship', enum: ['full-time', 'part-time', 'contract', 'temporary', 'volunteer', 'internship'] }, salary: { type: 'string', description: 'Filter by salary: 40k+, 60k+, 80k+, 100k+, 120k+, 140k+, 160k+, 180k+, 200k+' }, postedLimit: { type: 'string', description: 'Filter by post date: 24h, week, month', enum: ['24h', 'week', 'month'] }, experienceLevel: { type: 'string', description: 'Filter: internship, entry, associate, mid-senior, director, executive', enum: ['internship', 'entry', 'associate', 'mid-senior', 'director', 'executive'] }, industryId: { type: 'string', description: 'Filter by industry ID (comma-separated)' }, functionId: { type: 'string', description: 'Filter by job function ID (comma-separated)' }, under10Applicants: { type: 'boolean', description: 'Filter jobs with under 10 applicants' }, easyApply: { type: 'boolean', description: 'Filter Easy Apply jobs' }, page: { type: 'integer', description: 'Page number', default: 1 }, save_dir: { type: 'string', description: 'Directory to save cleaned JSON data' }, max_items: { type: 'integer', description: 'Maximum results (default: 10)', default: 10 }, }, required: [], },
• src/index.ts:391-417 (registration)

  Tool registration in the MCP server's listTools response, defining name, description, and inputSchema for search_jobs.
  { name: 'search_jobs', description: 'Search LinkedIn jobs. Returns cleaned data in TOON format.', inputSchema: { type: 'object', properties: { search: { type: 'string', description: 'Search jobs by title' }, companyId: { type: 'string', description: 'Filter by company ID' }, location: { type: 'string', description: 'Filter by location' }, geoId: { type: 'string', description: 'Filter by LinkedIn Geo ID' }, sortBy: { type: 'string', description: 'Sort by: relevance or date', enum: ['relevance', 'date'] }, workplaceType: { type: 'string', description: 'Filter: office, hybrid, remote', enum: ['office', 'hybrid', 'remote'] }, employmentType: { type: 'string', description: 'Filter: full-time, part-time, contract, temporary, volunteer, internship', enum: ['full-time', 'part-time', 'contract', 'temporary', 'volunteer', 'internship'] }, salary: { type: 'string', description: 'Filter by salary: 40k+, 60k+, 80k+, 100k+, 120k+, 140k+, 160k+, 180k+, 200k+' }, postedLimit: { type: 'string', description: 'Filter by post date: 24h, week, month', enum: ['24h', 'week', 'month'] }, experienceLevel: { type: 'string', description: 'Filter: internship, entry, associate, mid-senior, director, executive', enum: ['internship', 'entry', 'associate', 'mid-senior', 'director', 'executive'] }, industryId: { type: 'string', description: 'Filter by industry ID (comma-separated)' }, functionId: { type: 'string', description: 'Filter by job function ID (comma-separated)' }, under10Applicants: { type: 'boolean', description: 'Filter jobs with under 10 applicants' }, easyApply: { type: 'boolean', description: 'Filter Easy Apply jobs' }, page: { type: 'integer', description: 'Page number', default: 1 }, save_dir: { type: 'string', description: 'Directory to save cleaned JSON data' }, max_items: { type: 'integer', description: 'Maximum results (default: 10)', default: 10 }, }, required: [], }, } as Tool,
• src/index.ts:543-543 (registration)

  Dispatch case in the callTool request handler that routes search_jobs calls to the searchJobs method.
  case 'search_jobs': return await this.searchJobs(args as Record<string, any>);
• src/index.ts:119-130 (helper)

  Helper function in DataCleaners that cleans raw job search results, extracting key fields like id, title, url, company, location for agent-friendly output.
  cleanJobSearchResult(raw: any): any { if (!raw) return null; return { id: raw.id, title: raw.title, url: raw.url, postedDate: raw.postedDate, company: raw.company?.name, location: raw.location?.linkedinText, easyApply: raw.easyApply, }; },