mcpField
J
JobGPT

JobGPT

Local
6figr.comBusiness & Marketing26Updated 2d ago

About JobGPT

JobGPT MCP Server is the official MCP server for the JobGPT job search platform. It lets AI assistants in clients such as Claude, Cursor and Windsurf search jobs, manage applications and generate tailored resumes.

Key features

  1. 01Job search with multiple filters
  2. 02Job hunt and auto-apply management
  3. 03AI-tailored resume generation
  4. 04Application statistics and status tracking
  5. 05Recruiter lookup and outreach email drafts

Use cases

  1. 01Search for remote roles through an AI assistant
  2. 02Generate a tailored resume for a specific application
  3. 03Review recent application counts and statuses

MCP support

The project is an MCP server exposing 34 tools for job search, applications, resumes and outreach. Clients connect through browser sign-in with OAuth or an API key.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"jobgpt": {
"command": "npx",
"args": [
"-y",
"jobgpt-mcp-server"
],
"env": {
"JOBGPT_API_KEY": "<JOBGPT_API_KEY>"
}
}
}
}
Run commandnpx -y jobgpt-mcp-server
Environment variables to set
JOBGPT_API_KEYSecretYour JobGPT API key from 6figr.com/account (MCP Integrations section)

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npmjobgpt-mcp-server1.1.4

Tools

35

Saved tool list · updated 3w ago

ParameterTypeRequiredDescription
pagenumberNoPage number for pagination (default: 1)
limitnumberNoMaximum number of results (default: 5, max: 50). Keep low to avoid large responses.
remotebooleanNoFilter for remote jobs only
skillsarrayNoRequired skills (e.g., ["Python", "React", "AWS"])
titlesarrayNoJob titles to search for (e.g., ["Software Engineer", "Senior Developer"])
companiesarrayNoSpecific companies to search (e.g., ["Google", "Meta", "Apple"])
countriesarrayNoCountry codes to filter (e.g., ["US", "CA", "UK"])
expLevelsarrayNoExperience levels (e.g., ["SE" for Senior, "MI" for Mid-level, "EN" for Entry])
locationsarrayNoLocations to search in (e.g., ["San Francisco", "New York", "Remote"])
dateOffsetenumNoOnly show jobs posted within this time period (e.g., "2D" for last 2 days)24H | 1D | 2D | 7D | 14D | 1M | 3M | 6M | 9M | 1Y
industriesarrayNoFilter by company industries (use get_industries to see valid values)
companySizearrayNoFilter by company size (e.g., ["xs" for 1-50, "s" for 50-200, "m" for 200-1K, "l" for 1K-5K, "xl" for 5K+])
baseSalaryMaxnumberNoMaximum base salary (USD)
baseSalaryMinnumberNoMinimum base salary (USD)
h1bSponsorshipbooleanNoFilter for jobs offering H1B sponsorship
excludedCompaniesarrayNoCompanies to exclude from results

Details

From the project README

JobGPT MCP Server

The official MCP server for JobGPT — auto apply & search jobs, generate and manage custom tailored resumes, and track applications directly from Claude, Cursor, Windsurf, and any MCP-compatible AI tool.

JobGPT Server MCP server

What You Can Do

Ask your AI assistant things like:

  • "Find remote senior React jobs paying over $150k"
  • "Auto-apply to the top 5 matches from my job hunt"
  • "Generate a tailored resume for this Google application"
  • "Apply to this job for me - <job_url from company website, greenhouse workday or linkedin, etc>"
  • "Show my application stats for the last 7 days"
  • "Find recruiters for this job and draft an outreach email"

The MCP server connects your AI assistant to the full JobGPT platform — 34 tools covering job search, applications, resumes, outreach, and more.

Quick Start

There are two ways to connect, depending on your AI tool:

  • Browser login (recommended) — no API key, no JSON, no Node.js. Just add the server URL and sign in to 6figr in your browser. Works with Claude Desktop, Claude Code, Codex CLI (codex mcp login), and ChatGPT Web (Developer Mode).
  • API key — for the ChatGPT Desktop app, Cursor, Windsurf, Cline, Continue, and manual config. Paste a key into the tool's config.
Get an API Key (only for the API-key method)
  1. Go to 6figr.com/account
  2. Scroll to MCP Integrations
  3. Click Generate API Key
  4. Copy the key (starts with sk_)

The browser-login method does not need this — it fetches your key for you after you sign in.


Setup by Client

Claude Desktop

Recommended — browser login (no API key, no JSON):

  1. Open Settings → Connectors → Add custom connector
  2. Name it JobGPT and enter the URL https://mcp.6figr.com/mcp
  3. Click Connect — your browser opens a 6figr sign-in page
  4. Sign in and click Approve. Done — the JobGPT tools appear in Claude.

No terminal, no config file, no Node.js. This uses OAuth, so your API key never touches Claude's config — Claude gets a scoped token instead.

Alternative — API key via mcp-remote (for older Claude Desktop builds that don't show Connectors; requires Node.js 18+):

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "jobgpt": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.6figr.com/mcp",
        "--header",
        "Authorization:${AUTH_HEADER}"
      ],
      "env": {
        "AUTH_HEADER": "Bearer your-api-key-here"
      }
    }
  }
}
Claude Code (CLI)

Recommended — browser login (no API key): add the server, then authenticate in your browser:

claude mcp add jobgpt -t http -u https://mcp.6figr.com/mcp

Then run /mcp inside Claude Code, select jobgpt → Connect, and sign in to 6figr when the browser opens. Claude Code stores a scoped OAuth token — no API key needed.

Option A: One-liner with an API key (claude mcp add):

claude mcp add jobgpt -t http -u https://mcp.6figr.com/mcp --header "Authorization: your-api-key-here"

Option B: Edit settings.json manually

Add to ~/.claude/settings.json:

{
  "mcpServers": {
    "jobgpt": {
      "type": "http",
      "url": "https://mcp.6figr.com/mcp",
      "headers": {
        "Authorization": "your-api-key-here"
      }
    }
  }
}
Codex CLI (ChatGPT)

Recommended — browser login (no API key): add the server, then authenticate in your browser:

codex mcp add jobgpt --url https://mcp.6figr.com/mcp
codex mcp login jobgpt

codex mcp login opens a browser to sign in to 6figr and approve access; Codex stores a scoped OAuth token (OAuth 2.1 + PKCE). No API key needed.

Alternative — API key via header:

codex mcp add jobgpt --url https://mcp.6figr.com/mcp --header "Authorization: Bearer your-api-key-here"
ChatGPT

ChatGPT connects two different ways depending on where you add the server:

ChatGPT Desktop app — API key. Open Settings → Connectors → Advanced/Developer → Connect to a custom MCP and fill in:

  • Name: JobGPT
  • Type: Streamable HTTP
  • URL: https://mcp.6figr.com/mcp
  • Headers → Add header: Key Authorization, Value Bearer your-api-key-here

The desktop dialog authenticates by header (there's no browser-login button), so use your sk_ key here. Leave the env-var fields blank.

ChatGPT Web (Developer Mode) — browser login (OAuth). On the web app, enable Settings → Connectors → Advanced → Developer mode, then Add custom connector, name it JobGPT, and enter the URL https://mcp.6figr.com/mcp. ChatGPT runs the OAuth flow: sign in to 6figr in the popup and approve. No API key needed. (Developer mode is available on Plus, Pro, Business, Enterprise, and Edu plans.)

Cursor

Go to Settings > MCP > Add new MCP server, or add to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "jobgpt": {
      "type": "http",
      "url": "https://mcp.6figr.com/mcp",
      "headers": {
        "Authorization": "your-api-key-here"
      }
    }
  }
}
Windsurf

Go to Settings > Cascade > MCP > Add Server > Add custom server, or add to ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "jobgpt": {
      "type": "http",
      "url": "https://mcp.6figr.com/mcp",
      "headers": {
        "Authorization": "your-api-key-here"
      }
    }
  }
}
Cline (VS Code)

Open the Cline MCP settings in VS Code and add:

{
  "mcpServers": {
    "jobgpt": {
      "type": "http",
      "url": "https://mcp.6figr.com/mcp",
      "headers": {
        "Authorization": "your-api-key-here"
      }
    }
  }
}
Continue (VS Code / JetBrains)

Add to your Continue config (~/.continue/config.yaml):

mcpServers:
  - name: jobgpt
    type: http
    url: https://mcp.6figr.com/mcp
    headers:
      Authorization: "your-api-key-here"
Alternative: Run Locally

If you prefer to run the server on your machine (requires Node.js 18+):

{
  "mcpServers": {
    "jobgpt": {
      "command": "npx",
      "args": ["-y", "jobgpt-mcp-server"],
      "env": {
        "JOBGPT_API_KEY": "your-api-key-here"
      }
    }
  }
}

Available Tools

Job Search
ToolDescription
search_jobsSearch jobs with filters — titles, locations, companies, skills, salary, remote, H1B sponsorship
match_jobsGet new job matches from a saved job hunt (only unseen jobs)
get_jobGet full details of a specific job posting
Profile & Salary
ToolDescription
get_profileView your profile — skills, experience, work history, education
update_profileUpdate name, headline, location, skills, experience
get_salaryGet your current compensation details
update_salaryUpdate base salary, stocks, bonus, target salary
get_currenciesList supported currencies (for salary updates)
get_creditsCheck your remaining credits balance
Job Hunts
ToolDescription
list_job_huntsList your saved job hunts with credits balance
create_job_huntCreate a new job hunt with search filters and auto-apply settings
get_job_huntGet details of a specific job hunt
update_job_huntUpdate filters, auto-apply mode, daily limits, status
Applications
ToolDescription
get_application_statsAggregated stats — counts by status, auto-apply metrics
list_applicationsList applications filtered by job hunt or status
get_applicationGet full application details
update_applicationUpdate status or notes
apply_to_jobTrigger auto-apply for an application
add_job_to_applicationsSave a job from search results to your applications
import_job_by_urlImport a job from any URL (LinkedIn, Greenhouse, Lever, Workday, etc.)
Resume
ToolDescription
list_resumesList your uploaded resumes
get_resumeGet resume details and download URL
delete_resumeDelete an alternate resume
upload_resumeUpload a resume from URL (PDF, DOC, DOCX)
list_generated_resumesList AI-tailored resumes created for applications
get_generated_resumeGet a generated resume's download URL
generate_resume_for_jobGenerate an AI-optimized resume for a specific application
calculate_match_scoreCalculate resume-to-job match score with skill analysis
Outreach
ToolDescription
get_job_recruitersFind recruiters associated with a job
get_job_referrersFind potential referrers at a company
get_application_recruitersGet recruiters for a saved application
get_application_referrersFind referrers for a saved application
list_outreachesList your sent outreach emails
send_outreachSend an outreach email to a recruiter or referrer

Environment Variables

VariableRequiredDefaultDescription
JOBGPT_API_KEYYes—Your API key from 6figr.com/account
JOBGPT_API_URLNohttps://6figr.comAPI base URL
DEBUGNofalseEnable debug logging to stderr

Troubleshooting

"JOBGPT_API_KEY environment variable is required"

Your API key isn't being passed to the server. Make sure it's in the env block of your MCP config.

Tool calls failing with "API Error (401)"

Your API key is invalid or expired. Generate a new one at 6figr.com/account.

"You have run out of credits"

Some operations (auto-apply, resume generation) consume credits. Purchase more at 6figr.com/jobgpt.

Server not appearing in your AI tool
  1. Make sure Node.js 18+ is installed (node --version)
  2. Restart your AI tool after editing the config file
  3. Try running manually to check for errors: JOBGPT_API_KEY=your-key npx jobgpt-mcp-server
Debug mode

Add "DEBUG": "true" to your env config to see detailed API request/response logs in stderr.

Development

git clone https://github.com/6figr-com/jobgpt-mcp-server.git
cd jobgpt-mcp-server
npm install
cp .env.example .env   # add your API key

npm run dev:local       # run stdio server locally
npm run build           # compile TypeScript
npm run dev:worker      # run Cloudflare Worker locally
npm run deploy          # deploy to Cloudflare Workers
Testing with MCP Inspector
npx @modelcontextprotocol/inspector

License

MIT

Links

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.