Context Condenser MCP server

This is a small helper that trims down long code files and noisy logs before your AI reads them. It keeps the important parts, like function names and error messages, and drops the filler. It is handy if you work with big files and want your AI to stay fast, cheap, and focused.

What this MCP server does

You ask your AI something about a big code file or a long log. The AI passes the text to this helper. The helper reads the code and builds a short outline that keeps the imports, types, and function signatures but shrinks the bodies. For logs, it strips the noise and points at the real errors. Then the AI answers you using the short version, which uses fewer tokens and stays on track.

Click to zoom

What you can do with it

• Outline a whole code file into a short skeleton
• Pull out just one function, class, or type by name
• Compress a long test or build log down to the real failures
• Check how many tokens a file will cost before sending it
• Keep the file's imports and types when extracting a symbol
• See how many lines and tokens were saved

Try asking your AI

• “Outline src/services/auth.ts so I can see the structure”
• “Extract just the validateToken function from src/services/auth.ts”
• “Compress test_output.log and show me what actually failed”
• “How many tokens will src/large_file.ts use if I paste it in”

What it gives back to you

You get back a shorter version of what you sent. For code, that is a skeleton with the important names and signatures kept. For logs, you get the failing lines and error messages pulled out. It also tells you how many lines and tokens were saved.

Before you start

Who it's for

Developers and anyone who works with large code files or long logs and wants their AI to stay fast and cheap.