tools
Monorepo for Sylph Lab Model Context Protocol (MCP) tools and servers.
Documentation
π Overview
A comprehensive collection of modular tools designed for the Model Context Protocol (MCP), enabling AI agents to perform various operations efficiently.
The Problem:
Building MCP servers from scratch:
- Repetitive boilerplate code
- Inconsistent tool interfaces
- Scattered tool implementations
- No shared utilities or patternsThe Solution:
MCP Tools Monorepo:
- Modular tool architecture π§©
- Consistent interfaces across tools π―
- Shared core utilities π¦
- MCP adaptors for easy integration β‘Result: Build MCP servers faster with reusable, well-tested tool components.
π¦ Packages
Core & Infrastructure
| Package | Description | Purpose |
|---|---|---|
| tools-core | Core utilities and type definitions | Base framework for all tools |
| tools-adaptor-mcp | MCP server integration adaptor | Convert tools to MCP servers |
| tools-adaptor-vercel | Vercel platform adaptor | Deploy tools on Vercel |
Data Processing Tools
| Package | Description | MCP Server |
|---|---|---|
| tools-json | JSON parsing, validation, transformation | β tools-json-mcp |
| tools-xml | XML parsing and processing | β tools-xml-mcp |
| tools-base64 | Base64 encoding/decoding | β tools-base64-mcp |
| tools-pdf | PDF text extraction | β tools-pdf-mcp |
System & Network Tools
| Package | Description | MCP Server |
|---|---|---|
| tools-filesystem | File operations (read, write, list, search) | β tools-filesystem-mcp |
| tools-net | Network requests (fetch, IP info, DNS) | β tools-net-mcp |
| tools-fetch-mcp | Specialized fetch tool for MCP | β Standalone MCP |
| tools-hasher | Hashing utilities (MD5, SHA, etc.) | β tools-hasher-mcp |
Specialized Tools
| Package | Description | MCP Server |
|---|---|---|
| tools-rag | Retrieval-Augmented Generation tools | β tools-rag-mcp |
| tools-wait | Delay and timing utilities | β tools-wait-mcp |
Pattern: Most tools follow the `tools-{name}` + `tools-{name}-mcp` pattern, separating core logic from MCP server implementation.
π― Use Cases
AI Agent Development
Build powerful MCP servers for AI assistants:
- File management - Read, write, search files
- Web scraping - Fetch and process web content
- Document processing - Extract text from PDFs
- Data transformation - JSON/XML/Base64 operations
Rapid Prototyping
Quickly assemble MCP servers from pre-built tools:
- Mix and match - Combine tools as needed
- Consistent interfaces - Predictable tool behavior
- Shared utilities - Less code duplication
Production Deployment
Deploy robust MCP servers:
- Vercel adaptor - Serverless deployment
- Turborepo - Optimized builds
- Type-safe - Full TypeScript support
π Quick Start
Prerequisites
- Node.js - Latest LTS version (check `.nvmrc` if available)
- pnpm - Version specified in `package.json` `packageManager` field
Installation
# Clone the repository
git clone https://github.com/SylphxAI/tools.git
cd tools
# Install dependencies
pnpm installDevelopment Workflow
# Build all packages
pnpm build
# Build continuously during development
pnpm build:watch
# Run linting
pnpm lint
# Run tests (build first)
pnpm testπ οΈ Tech Stack
| Component | Technology |
|---|---|
| Monorepo | Turborepo |
| Package Manager | pnpm workspaces |
| Language | TypeScript |
| Code Quality | Biome (50x faster than ESLint) |
| Testing | Vitest |
| MCP Protocol | Model Context Protocol |
π‘ Architecture
Tool Structure
Each tool follows a consistent pattern:
packages/
βββ tools-{name}/ # Core tool logic
β βββ src/
β β βββ index.ts # Tool implementation
β β βββ types.ts # TypeScript types
β βββ package.json
βββ tools-{name}-mcp/ # MCP server wrapper
βββ src/
β βββ index.ts # MCP server entry
βββ package.jsonBenefits
- Separation of concerns - Core logic separate from MCP protocol
- Reusability - Use tools outside of MCP context
- Testability - Test core logic independently
- Flexibility - Easy to create alternative adaptors
π Package Details
Core Utilities
tools-core
import { defineTool } from '@sylphx/tools-core'
const myTool = defineTool({
name: 'my-tool',
description: 'Does something useful',
schema: { /* Zod schema */ },
execute: async (input) => { /* implementation */ }
})MCP Adaptor
tools-adaptor-mcp
import { createMCPServer } from '@sylphx/tools-adaptor-mcp'
import { myTool } from '@sylphx/tools-mytool'
const server = createMCPServer({
tools: [myTool],
name: 'My MCP Server',
version: '1.0.0'
})π§ Development
Building Packages
# Build all packages
pnpm build
# Build specific package
pnpm --filter @sylphx/tools-filesystem build
# Watch mode for development
pnpm build:watchCode Quality
# Lint all packages
pnpm lint
# Fix linting issues
pnpm lint:fix
# Format code
pnpm formatTesting
# Run all tests
pnpm test
# Run tests for specific package
pnpm --filter @sylphx/tools-json test
# Watch mode
pnpm test:watchπ Creating a New Tool
Step 1: Create Core Package
# Create package directory
mkdir -p packages/tools-newtool/src
# Create package.json
cat > packages/tools-newtool/package.json {
// Implementation
return { result: `Processed: ${input}` }
}
})Step 3: Create MCP Server
# Create MCP package
mkdir -p packages/tools-newtool-mcp/src
# Create server
cat > packages/tools-newtool-mcp/src/index.ts
Modular. Reusable. Production-ready.
The MCP toolkit that scales with your needs
β’
β’Frequently asked questions
What is tools?
tools is Monorepo for Sylph Lab Model Context Protocol (MCP) tools and servers.
How do I install tools?
Open the GitHub repository and follow its README. Most MCP servers are added to your client's MCP config, then called by your agent.
Is tools open source?
Yes β it is hosted on GitHub at https://github.com/sylphlab/MCP and has 4 stars.
Related MCP tools
A Model Context Protocol server for converting almost anything to Markdown TypeScript-based implementation. Trusted by 2200+ developers.
Composio equips your AI agents & LLMs with 100+ high-quality integrations via function calling for the Model Context Protocol. Enhance AI assistants with powerf
π¦ Repomix is a powerful tool that packs your entire repository into a single, AI-friendly file. Perfect for when you need to feed your codebase to Large Lan...
The TypeScript AI agent framework. β‘ Assistants, RAG, observability. Supports any LLM: GPT-4, Claude, Gemini, Llama. Built for the Model Context Protocol to enh
Rich-text editor with AI, MCP, and shadcn/ui TypeScript-based implementation. Trusted by 15300+ developers. Trusted by 15300+ developers.
MCP server to provide Figma layout information to AI coding agents like Cursor TypeScript-based implementation. Trusted by 11600+ developers.
Run your own MCP server? See who uses it and what to fix.
Measure it with TrackMCP