MCP server · Developer tools

tokentoll MCP server

Lets your AI scan your code for LLM API calls and estimate what they cost per month.

Developer tools For it and development Developer Tools

tokentoll is a small tool that reads your code and finds every place you call an AI model, like OpenAI or Claude. It then estimates how much those calls would cost per month. This MCP server connects that tool to your AI chat, so you can ask your AI about cost questions instead of running commands yourself.

What is an MCP server? The 30-second version

On its own, your AI can only chat. An MCP server is a small helper program that gives your AI a new skill. This one gives your AI the ability to run tokentoll, which looks through your code files and turns AI calls into dollar estimates. Once it is connected, you can just ask your AI to check costs and it will do it for you.

What this MCP server does

You ask your AI something like how much a folder of code costs per month. Your AI passes that request to the tokentoll helper. The helper reads your Python, JavaScript, or TypeScript files, finds AI model calls (OpenAI, Anthropic, Google, LangChain, and more), and looks up prices from a bundled price list. It then sends back an estimate. If you ask it to compare two versions of your code, it shows the cost difference between them.

What you can do with it

Scan a folder of code and list every AI model call it finds
Estimate the monthly cost of those calls in dollars
Compare the cost between two versions of your code, like before and after a change
Spot a call that got much more expensive after an edit
Check whether a code change uses a model with no known price

Try asking your AI

“Scan my src folder and tell me the estimated monthly cost”
“Compare the cost between my main branch and this branch”
“Which AI calls in this project would cost the most per month”
“Did my last change make any AI call more expensive”

What it gives back to you

You get back a summary in the chat: a list of the AI calls found, the model each one uses, and an estimated monthly cost in dollars. If you asked for a comparison, you also see the difference between the two versions, like plus 800 dollars or minus 12 dollars. When a call is more expensive than before, it is flagged so you can look closer. Everything comes back as text and numbers you can read right in the conversation.

Before you start

What you need

Python installed on your computer (version 3.10 or newer)
A folder of code with AI API calls, like Python or TypeScript files
The tokentoll package installed with its MCP extras

Good to know

The cost numbers are estimates based on assumed monthly call volume and a simple token guess, so treat them as rough guidance, not exact bills.

Install it with your AI

Add tokentoll MCP server to your AI, no technical skills needed

You don't install anything by hand. You copy one prompt, paste it into an AI that can work on your computer, and it checks, installs and connects the server for you, asking you when it needs something.

Sign in to get the install prompt

Members get a ready-made prompt that lets the Claude desktop app check tokentoll MCP server, install it and connect it for them, step by step. You don't need any technical skills: you copy, paste and answer a few questions. Your connected AI can also find and install any of the 4,609 MCP servers here for you.