worklittle-jobs-mcp

Worklittle Jobs MCP

Connect AI assistants to Worklittle for job search, application, resume, and cover letter generation across 4+ million job listings.

Source

• GitHub: https://github.com/worklittle/jobs-mcp
• Official Registry: io.github.worklittle/jobs
• Remote MCP: https://mcp.worklittle.com/
• Docs: https://docs.worklittle.com/mcp
• Support: support@worklittle.com

What It Does

• Search 4M+ jobs with filters: visa sponsorship, salary, distance, location, company, job type, seniority
• Swipe & apply — connect your Worklittle account to save jobs and apply directly
• Generate documents — create hiring-manager-ready resumes and cover letters
• Market insights — get aggregate job market statistics and trends
• ATS tools — manage job listings and candidates (employer-side)

Installation

Add MCP Server

In your Hermes Agent config (config.yaml), add:

mcpServers:
  worklittle:
    url: https://mcp.worklittle.com/
    headers:
      Authorization: "Bearer sk-wl-api01-YOUR_API_KEY"

Or run via CLI:

hermes mcp add worklittle --url https://mcp.worklittle.com/

Get API Key

1. Sign up at https://worklittle.com/business
2. Generate API key at https://worklittle.com/business/api-keys
3. Keys start with sk-wl-api01-

Tools

Job Search & Application

Document Generation

Account & Profile

ATS (Employer Tools)

Usage

Search Jobs

# Example: search remote software engineer jobs with visa sponsorship
search_jobs(
    query="software engineer",
    location="remote",
    visa_sponsored=True,
    min_salary=120000,
    distance_miles=50,
    limit=20
)

Get Job Details

# After searching, get details for a specific role
get_job_details(job_id="abc123")

Generate Keywords for Resume Tailoring

# Extract skills from a job description (better than using full JD text)
keywords = get_job_keywords(job_id="abc123")
# Returns: ["Python", "React", "AWS", "GraphQL", "Kubernetes", ...]

Create Resume

# One-page resume from your context
resume = create_resume(
    format="pdf",  # html, pdf, or docx
    context={"experience": [...], "education": [...]}
)

Apply to a Job

# Submit application (requires jobs:apply scope on API key)
submit_job_application(
    job_id="abc123",
    resume_url="...",
    cover_letter="..."
)
# Cost: $0.01 per successful submission

Market Overview

# Get high-level job market stats
overview = get_market_overview()
# Returns: total_jobs, remote_ratio, top_hiring_companies, salary_ranges...

Prompt Guide

Job Search Flow

1. User: "Find me remote Python jobs with visa sponsorship paying $150k+"
2. Agent: Call search_jobs with filters → present results with salary, visa status, distance
3. User: "Tell me more about X"
4. Agent: Call get_job_details(job_id) → summarize key requirements
5. User: "What skills does this role need?"
6. Agent: Call get_job_keywords(job_id) → show distilled tech stack
7. User: "Apply to this one"
8. Agent: Call submit_job_application or open Jobs app via search_jobs app mode

Resume Building Flow

1. User: "Tailor my resume for this job"
2. Agent: Call get_job_keywords(job_id) to get exact skill requirements
3. Agent: Call create_resume with job context → generates targeted resume
4. Agent: Suggest specific improvements based on keyword gap analysis

Protocol

• Transport: JSON-RPC 2.0 over HTTP (Streamable HTTP)
• Auth: Bearer token (sk-wl-api01-*)
• Base URL: https://mcp.worklittle.com/
• Dual-era support: Modern 2026-07-28 Streamable HTTP + legacy initialize protocols

Pitfalls

• API key scope matters — submit_job_application requires jobs:apply scope
• External vs Worklittle jobs — submit_job_application only works for employer-posted jobs on Worklittle, not external apply_url listings
• Keyword vs full JD — use get_job_keywords instead of full job description for resume/cover letter generation (better quality)
• No vector embeddings — search is substring/FTS based, not semantic
• Negative terms — use leading dash to exclude, e.g., "software engineer -junior"
• Price per application — $0.01 per successful submit; budget accordingly for bulk applications

Integration with Other Tools

# Combined workflow: search → apply → track
# Use with:
# - hermes mcp tools (native)
# - mcporter CLI (ad-hoc)
# - REST API directly (for apps without MCP libraries)

Verification

After installing, test with:

# Check MCP connection
hermes mcp list

# Test search
hermes mcp call worklittle search_jobs '{"query": "remote python", "limit": 3}'

# Expected: JSON array of job listings with title, company, salary, location

References

• GitHub: https://github.com/worklittle/jobs-mcp
• Documentation: https://docs.worklittle.com/mcp
• Official Registry: io.github.worklittle/jobs
• Privacy: https://worklittle.com/privacy