Terse MCP server

Terse is a small programming language made for AI coding agents rather than people. This MCP server lets your AI write Terse code, run it, turn it into Python, and count how many tokens it uses. It is handy if you are curious about AI-written code or want a lighter way for your assistant to do small computing tasks.

What this MCP server does

You ask your AI to do something with Terse, like run a small program or check a snippet. The AI writes the Terse code and hands it to this helper program on your computer. The helper reads the code, runs it in a safe way, and sends back the result. If something is wrong, it sends back a clear error message with a code and a hint, so the AI can fix it and try again. You see the answer, the printed output, or the translated Python right in your chat.

What you can do with it

• Run a small Terse program and see its output
• Turn Terse code into readable Python
• Check Terse code for mistakes before running it
• Count how many tokens a piece of Terse code uses
• Parse Terse code into a structured form your AI can inspect
• Try out small computing tasks like fibonacci or fizzbuzz in Terse

Try asking your AI

• “Run this Terse program and tell me what it prints: (print (add 40 2))”
• “Translate this Terse code into Python: (def fib(n)(if(lt n 2)n(add(fib(dec n))(fib(dec(dec n))))))”
• “Check this Terse snippet for errors: (add 1)”
• “How many tokens does this Terse code use compared to the same thing in Python?”

What it gives back to you

You get back the result of whatever you asked for. For a run, that is the value and any printed output. For a check, you get either OK or a short error message with a code and a hint. For a transpile, you get Python source code you can read. For a token count, you get a number.

Before you start

Who it's for

People who are curious about AI-written code, developers experimenting with small languages, and anyone who wants a lighter way for their assistant to do quick computing tasks.