TokenScope: Token-Aware Directory Explorer MCP Server

@cdgaete

代码仓库

基本信息

分类

文件与存储

传输方式

stdio

作者

cdgaete

提交者

Carlos Gaete

关于 TokenScope: Token-Aware Directory Explorer MCP Server

A tool to analyze directory structures while being mindful of token limitations for LLMs. Customize ignore files like gitignore patterns, such as .venv or node_modules

#mcp #directory #explorer #claude #ai #assistant#token-aware

TokenScope: Token-Aware Directory Explorer MCP Server 配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "TokenScope": {
      "command": "uv",
      "args": [
        "run",
        "--with",
        "tokenscope",
        "tokenscope",
        "--base-path",
        "/your/secure/base/path"
      ]
    }
  }
}

TokenScope: Token-Aware Directory Explorer MCP Server 工具

工具列表暂未提供

当前收录暂时没有可展示的工具列表，不代表服务没有工具。请查阅项目文档了解支持的能力。

工具是从 README 中自动提取的。维护者可以在 ## Tools 标题下列出工具,即可填充这部分内容。

概览

What is TokenScope: Token-Aware Directory Explorer?

TokenScope is a Model Context Protocol (MCP) server that provides token-aware directory structure analysis and file content extraction for Large Language Models. It helps LLMs understand codebases efficiently by respecting token limits and filtering via .gitignore patterns.

How to use TokenScope: Token-Aware Directory Explorer?

Configure TokenScope with a mandatory base path for security, then invoke its six tools (e.g., scan_directory_structure, extract_file_content, search_files_by_pattern, analyze_token_usage, generate_directory_report, copy_file_to_destination) from any MCP-compatible client. All operations are confined to the allowed directory and automatically respect .gitignore rules.

Key features of TokenScope: Token-Aware Directory Explorer

Recursive directory scanning with token-efficient summaries
Smart file content extraction respecting token limits
Pattern-based file search within the allowed directory
Token usage estimation for entire directories
Comprehensive directory reports with structure and token statistics
Secure file copying between locations inside the base path

Use cases of TokenScope: Token-Aware Directory Explorer

Give an LLM structured context about a codebase before code generation
Extract relevant source files from a large project while staying under token budgets
Estimate the token cost of processing a directory with an LLM
Quickly locate files matching specific patterns in a project
Generate token-aware overviews of directory contents for AI-assisted development

常见问题

What is the base path and why is it required?

The base path is a mandatory directory that limits all file operations. It ensures security by confining scans, extractions, and copies to that directory only.

How does TokenScope handle token limits?

The server estimates token counts and uses smart extraction to stay within configurable token limits, providing only as much content as the LLM can process.

Does TokenScope respect .gitignore patterns?

Yes, it automatically filters out files and directories matched by .gitignore patterns in the scanned directory, reducing noise and saving tokens.

What are the runtime dependencies of TokenScope?

TokenScope is an MCP server; it requires a runtime environment with the MCP protocol and access to the filesystem it scans. No additional databases or external services are needed.

How secure is file copy in TokenScope?

The `copy_file_to_destination` tool operates exclusively within the allowed base path, preventing any file operations outside that directory. Destination files must also reside inside the base path.

文件与存储分类下的更多 MCP 服务器

Storacha MCP Storage Server

storacha

Storacha MCP storage server - self-sovereign data for your AI applications.

PDF Reader MCP Server (@shtse8/pdf-reader-mcp)

shtse8

📄 The PDF intelligence layer for AI agents — Agent Document Twin, evidence-first extraction, visual crops, OCR provenance, trust reports, and benchmark-gated releases. MCP server for Claude, Cursor, VS Code, and any MCP client.

810

drop.media

Upload once. Find, use and share across your AI workflows. drop.media is one media library for you and every AI agent you use: files added from Claude, ChatGPT, Cursor, the terminal or the web can be found, reused and sh

Filesystem

modelcontextprotocol

Model Context Protocol Servers

88.1K

File Context Server

bsmi021

A Model Context Protocol (MCP) server that provides file system context to Large Language Models (LLMs). This server enables LLMs to read, search, and analyze code files with advanced caching and real-time file watching capabilities.