Carrerlift MCP

Fresh Indian jobs and internships, plus international new-grad and intern roles, inside any AI assistant that speaks the Model Context Protocol.

Ask Claude, ChatGPT, Cursor or VS Code things like:

"Find React internships in Bangalore posted this week."
"Any remote data analyst roles for freshers?"
"Show me ML internships in the USA at big tech companies."
"Open this listing and tell me how to apply: carrerlift.in/jobs/…"

The assistant searches carrerlift.in live and answers with real listings, each linked to its page.

Endpoint

Code

https://www.carrerlift.in/api/mcp

It's free, needs no sign-up or API key, and is read-only.

Tools

Tool	What it does
`search_jobs`	Searches live jobs and internships in India, newest first. You can filter by keywords (role, skill or company), city and type (Internship, Full-time, Full-time Internship, Apprenticeship). Returns up to 24 per page.
`get_job`	Returns one listing in full: the complete description, pay, and how to apply (apply link, HR email or WhatsApp, whichever the employer gave). Takes a slug or a `carrerlift.in/jobs/…` link.
`search_global_jobs`	Searches internships and new-grad roles abroad, mostly in the USA. The feed is merged from community-maintained GitHub job lists and refreshed a few times a day. You can filter by `intern_usa`, `newgrad_usa`, `intern_intl`, `newgrad_intl`, and by field (Software, AI/ML/Data, Hardware, Quant, Product, Other).
`list_job_filters`	Lists the cities and job types the Indian listings currently cover.

Indian listings are refreshed every 30 minutes. Closed listings with an apply link are left out.

Connect

Claude (claude.ai and the desktop app)

Go to Settings → Connectors → Add custom connector.
Enter the name Carrerlift and the URL https://www.carrerlift.in/api/mcp.
Enable it in a chat from the tools menu.

Claude Code

Terminal

claude mcp add --transport http carrerlift https://www.carrerlift.in/api/mcp

ChatGPT

Turn on Settings → Apps & Connectors → Advanced → Developer mode.
Under Settings → Apps & Connectors → Create, add https://www.carrerlift.in/api/mcp and set authentication to No authentication.

Cursor

Add this to ~/.cursor/mcp.json:

config.json

{
  "mcpServers": {
    "carrerlift": {
      "url": "https://www.carrerlift.in/api/mcp"
    }
  }
}

VS Code

Add this to .vscode/mcp.json:

config.json

{
  "servers": {
    "carrerlift": {
      "type": "http",
      "url": "https://www.carrerlift.in/api/mcp"
    }
  }
}

Clients that only run local (stdio) servers

Bridge to the remote endpoint with mcp-remote:

config.json

{
  "mcpServers": {
    "carrerlift": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://www.carrerlift.in/api/mcp"]
    }
  }
}

Try it with curl

Terminal

curl -s https://www.carrerlift.in/api/mcp \
  -H 'content-type: application/json' \
  -H 'accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"search_jobs","arguments":{"query":"python","location":"Pune"}}}'

Details

Transport: Streamable HTTP. The server is stateless and answers in JSON (no sessions, no SSE stream). GET returns 405.
Privacy: there's no login and no personal data. The server doesn't ask for or store anything about you. Links carry a utm_source=mcp tag so we can count visits that come from assistants.
Limits: results are paged, and paging stops at page 20. Please use the filters instead of crawling.
Where the data comes from: listings are collected and cleaned by Carrerlift. The international feed credits its source lists on carrerlift.in/global.

Feedback

Found a bug or have a tool request? Open an issue, or write to contactcarrerlift@gmail.com.

This repository holds the documentation and the MCP Registry manifest (server.json). The server runs as part of carrerlift.in.

License

MIT