mcptokens
Provides tool to count token cost of any MCP server's tool definitions, enabling agents to assess cost before enabling.
README
mcptokens
505 tokens in your agent's harness. One tool, inspect, that
counts the tool-definition cost of any other MCP server (stdio or
Streamable HTTP) before you enable it.
pip install mcptokens
<p> <a href="https://pypi.org/project/mcptokens/"><img src="https://img.shields.io/pypi/v/mcptokens.svg" alt="PyPI"></a> <a href="LICENSE"><img src="https://img.shields.io/badge/license-MIT-yellow.svg" alt="License: MIT"></a> <a href="https://www.python.org/downloads/"><img src="https://img.shields.io/badge/python-3.11+-blue.svg" alt="Python 3.11+"></a> <a href="https://github.com/dondai1234/mcptokens"><img src="https://img.shields.io/badge/Repo-dondai1234%2Fmcptokens-1f6feb.svg" alt="Repo"></a> </p>
Why
When your agent enables a new MCP server, its tools enter the
agent's context on every turn. The cost compounds fast. mcptokens
lets the agent ask is this server cheap enough? before flipping
the switch.
Transports
Same tool, same response shape:
- stdio (default): spawn a local process and speak JSON-RPC on its stdin/stdout.
- streamable_http: POST
initializeandtools/listto a remote MCP endpoint per MCP 2025-03-26. Server may reply viaapplication/json(one message) ortext/event-stream.
How
Add mcptokens to your agent's MCP config (Claude Code, Pi,
OpenCode, Codex, ...). The agent gains one tool:
inspect(command=["python", "-m", "some_mcp_server"])
# stdio spawn argv, same as your MCP config
inspect(command=["hound"])
# pre-installed binary
inspect(
transport="streamable_http",
url="http://localhost:8080/mcp",
headers={"Authorization": "Bearer ..."}, # optional
)
# remote MCP server
# Returns the same JSON shape every call:
# {ok, server, tool_count, wire_total_tokens,
# tools: [{name, total}], encoding, elapsed_ms, version}
wire_total_tokens is the number to report. Use it BEFORE
enabling a candidate server: a large value means don't enable.
The numbers
| Self-cost on wire | 505 tokens of cl100k_base |
| Tools exposed | 1, named inspect |
| Transports | stdio, streamable_http |
| Cross-platform | Linux, macOS, Windows |
| Python | 3.11+ |
| Imports | stdlib, tiktoken, mcp |
Install
pip install mcptokens
Add mcptokens to your agent's mcpServers / mcp_servers. Done.
CLI (debug surface)
mcptokens python -m some_mcp_server # human table
mcptokens --json python -m some_mcp_server # pipeline-friendly JSON
mcptokens --timeout 30 python -m srv # custom spawn, custom timeout
mcptokens serve # run as an MCP server
| Flag | Default | Purpose |
|---|---|---|
--encoding |
cl100k_base |
cl100k_base or o200k_base |
--timeout |
15 |
Per-server timeout in seconds (1 to 60) |
--json |
false |
Output JSON instead of the table |
License
MIT. pip install mcptokens from the canonical PyPI index.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.