Open source · PyPI

Give your AI a run button.

hackerearth-mcp plugs into Claude Desktop (or any MCP client) and adds one tool: run_code. Your AI writes code, runs it in HackerEarth's sandbox, sees the output, fixes its mistakes. 24 languages. No setup beyond a free API key.

Install from PyPI GitHub

24languages

2tools, nothing more

1000free runs / month

Two tools. That's the whole API.

MCP servers should be boring. This one does exactly two things, well.

// run_code

Execute anything

Send source code, get stdout, stderr, and timing back. Python to Rust to Haskell.

24 languages via HackerEarth's sandbox
Your AI can iterate: run, read error, fix, re-run
1000 free runs + 1000 compiles a month

// list_languages

Know what's available

Ask what languages the sandbox supports before writing a line.

Version info for every language
No guessing, no failed runs

// five-minute-setup

Install in one command

pip install, paste your free HackerEarth key, add one JSON block to Claude Desktop.

pip install hackerearth-mcp
Free key from hackerearth.com in 2 minutes
Published on the official MCP registry

How it works

From zero to running code inside your AI in three steps.

Install

pip install hackerearth-mcp

Key

Grab a free HackerEarth API key. Set it once.

Plug in

Add the server to Claude Desktop's config. Ask it to run something.