How many tokens does an MCP server cost, and how to cut it

Measure the tool list

A remote server answers tools/list over plain HTTP. Add an Authorization header if it needs a key. A server that insists on a session needs an initialize call first:

curl -s https://example.com/mcp \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

Four characters a token is an estimate. Each model has its own tokenizer and each client wraps tools its own way, so use the number to compare servers and track changes. In Claude Code, /context shows what is using your context window.

When the cost applies

Claude Code defers tool definitions by default through tool search: only tool names and server instructions load at session start, and Claude pulls in a tool's full definition when it looks that tool up. Tools load upfront when tool search is off or the deployment doesn't support it. Claude Code also cuts a description loaded upfront at 4,096 characters, so put what matters first.

Other clients handle this differently, so check the docs for yours. Your own server's size still matters under deferral: every definition the agent does load counts, and a client without deferral loads them all.

What to trim first

• Repeated regexes. zod's .uuid() writes a long pattern into every id field. Check the id with a custom validator instead and keep format: "uuid" in the schema.
• Integer bounds of 9007199254740991. Schema libraries add these safe-integer limits to integers, and they tell the model nothing. Set a real maximum or drop them.
• $schema on every tool, if your SDK lets you strip it.
• Descriptions that explain what the name already says. Keep what changes which tool the model picks and how it calls it.
• Tools with many input fields. Split them, or drop fields people rarely use.

Rules of thumb we use are about 5,000 tokens a server, 800 a tool and eight input fields per tool. No client enforces those numbers.

What we found on our own server

Taskhold's server has 22 tools and measured about 8,000 tokens. All 20 id fields carried zod's 180-character regex. A custom check kept the same validation and error message, and the list dropped to about 7,200.

We're still over our own rule of thumb, on purpose for now: the remaining descriptions are what tell an agent how to report progress and how to ask you a question. A test in our build fails if the list grows, so the number can only go down. See how agents report into Taskhold for what those tools do.