Mcp Tool Factory

MCP Tool Factory (TypeScript)

Generate production-ready MCP (Model Context Protocol) servers from natural language descriptions, OpenAPI specs, database schemas, GraphQL schemas, or ontologies.

[npm version] [npm downloads] [CI] [License: MIT] [TypeScript] [Node.js] [MCP]

Why MCP?

The Model Context Protocol (MCP) is an open standard that enables AI assistants to securely connect with external data sources and tools. MCP servers expose tools that can be used by:

• Claude Code and Claude Desktop
• OpenAI Agents SDK
• Google ADK (Agent Development Kit)
• LangChain and CrewAI
• Any MCP-compatible client

MCP Tool Factory lets you generate complete, production-ready MCP servers in seconds.

Features

Use as MCP Server

MCP Tool Factory is itself an MCP server! Add it to Claude Desktop, Claude Code, Cursor, or VS Code to generate MCP servers through conversation.

Tier 1 — Zero Config (Claude Code)

Claude Code auto-injects CLAUDE_CODE_OAUTH_TOKEN — no env vars needed:

claude mcp add mcp-tool-factory -- node /path/to/mcp-tool-factory-ts/bin/mcp-server.js

Tier 2 — Standard (Pick a Provider)

Set one API key and go. The factory auto-detects the provider:

Claude Desktop / Cursor / VS Code — add to your MCP config (claude_desktop_config.json, .cursor/mcp.json, or .vscode/mcp.json):

{  "mcpServers": {    "mcp-tool-factory": {      "command": "node",      "args": ["/path/to/mcp-tool-factory-ts/bin/mcp-server.js"],      "env": {        "ANTHROPIC_API_KEY": "your-key-here"      }    }  }}

Any of these API keys will work: ANTHROPIC_API_KEY, OPENAI_API_KEY, GOOGLE_API_KEY, MISTRAL_API_KEY, DEEPSEEK_API_KEY, GROQ_API_KEY, XAI_API_KEY, AZURE_OPENAI_API_KEY, COHERE_API_KEY.

Tier 3 — Full Control (Provider + Model + Budget)

Use MCP_FACTORY_PROVIDER, MCP_FACTORY_MODEL, and MCP_FACTORY_BUDGET to override auto-detection:

{  "mcpServers": {    "mcp-tool-factory": {      "command": "node",      "args": ["/path/to/mcp-tool-factory-ts/bin/mcp-server.js"],      "env": {        "OPENAI_API_KEY": "your-key-here",        "MCP_FACTORY_PROVIDER": "openai",        "MCP_FACTORY_MODEL": "gpt-5.2",        "MCP_FACTORY_BUDGET": "0.50"      }    }  }}

Claude Code CLI with full control:

claude mcp add mcp-tool-factory \  -e DEEPSEEK_API_KEY=your-key \  -e MCP_FACTORY_PROVIDER=deepseek \  -e MCP_FACTORY_MODEL=deepseek-chat \  -e MCP_FACTORY_BUDGET=0.25 \  -- node /path/to/mcp-tool-factory-ts/bin/mcp-server.js

Available Tools

Example Conversation

You: Create an MCP server for the GitHub API with tools to list repos, create issues, and manage pull requests

Claude: Uses generate_mcp_server tool

I've generated a complete MCP server with the following tools:

• list_repositories - List user repositories
• create_issue - Create a new issue
• list_pull_requests - List PRs for a repo
• merge_pull_request - Merge a PR

Let me write these files to your project...

Quick Start

Installation

# Global installationnpm install -g @heshamfsalama/mcp-tool-factory
# Or use npxnpx @heshamfsalama/mcp-tool-factory generate "Create tools for managing a todo list"

Set Your API Key

At least one provider API key is required:

# Anthropic Claude (recommended)export ANTHROPIC_API_KEY=your-key-here
# Or Claude Code OAuthexport CLAUDE_CODE_OAUTH_TOKEN=your-token-here
# Or any other supported providerexport OPENAI_API_KEY=your-key-hereexport GOOGLE_API_KEY=your-key-hereexport MISTRAL_API_KEY=your-key-hereexport DEEPSEEK_API_KEY=your-key-hereexport GROQ_API_KEY=your-key-hereexport XAI_API_KEY=your-key-hereexport AZURE_OPENAI_API_KEY=your-key-hereexport COHERE_API_KEY=your-key-here

Generate Your First Server

# From natural languagemcp-factory generate "Create tools for fetching weather data by city and converting temperatures"
# From OpenAPI specmcp-factory from-openapi ./api-spec.yaml
# From databasemcp-factory from-database ./data.db
# From GraphQL schemamcp-factory from-graphql ./schema.graphql
# From ontologymcp-factory from-ontology ./ontology.owl --format rdf

Usage

Natural Language Generation

mcp-factory generate "Create tools for managing a todo list with priorities" \  --name todo-server \  --output ./servers/todo \  --web-search \  --logging \  --metrics

OpenAPI Specification

# From local filemcp-factory from-openapi ./openapi.yaml --name my-api-server
# With custom base URLmcp-factory from-openapi ./spec.json --base-url https://api.example.com

Database Schema

# SQLitemcp-factory from-database ./myapp.db --tables users,posts,comments
# PostgreSQLmcp-factory from-database "postgresql://user:pass@localhost/mydb" --type postgresql

GraphQL Schema

# From a GraphQL SDL filemcp-factory from-graphql ./schema.graphql --name my-graphql-server
# From a URL endpointmcp-factory from-graphql https://api.example.com/graphql --name my-api-server

GraphQL queries are mapped to read-only MCP tools, and mutations are mapped to write tools. GraphQL types are automatically converted to Zod validation schemas.

Ontology

# From RDF/OWL (.owl, .rdf, .ttl)mcp-factory from-ontology ./ontology.owl --format rdf --name knowledge-server
# From JSON-LD (.jsonld)mcp-factory from-ontology ./schema.jsonld --format jsonld --name linked-data-server
# From custom YAML ontologymcp-factory from-ontology ./domain.yaml --format yaml --name domain-server

OWL Classes are mapped to MCP Resources, ObjectProperties become Tools, and DataProperties become tool parameters.

Test & Serve

# Run testsmcp-factory test ./servers/my-server
# Start server for testingmcp-factory serve ./servers/my-server

Generated Server Structure

servers/my-server/├── src/│   └── index.ts          # MCP server with tools, resources, and prompts├── tests/│   └── tools.test.ts     # Vitest tests (InMemoryTransport)├── package.json          # Dependencies├── tsconfig.json         # TypeScript config├── Dockerfile            # Container deployment├── README.md             # Usage documentation├── skill.md              # Claude Code skill file├── server.json           # MCP Registry manifest├── EXECUTION_LOG.md      # Generation trace (optional)└── .github/    └── workflows/        └── ci.yml        # GitHub Actions CI/CD

Generated servers export a createServer() factory function for easy testing. The server uses Streamable HTTP transport with a single /mcp POST endpoint and a /health GET endpoint. Tests use InMemoryTransport.createLinkedPair() for fast, reliable in-process testing with vitest.

CLI Reference

Generate Options

mcp-factory generate "..." \  --output, -o <path>           # Output directory (default: ./servers)  --name, -n <name>             # Server name  --description, -d <desc>      # Package description  --github-username, -g <user>  # GitHub username for MCP Registry  --version, -v <ver>           # Server version (default: 1.0.0)  --provider, -p <provider>     # LLM provider (anthropic, openai, google, mistral, deepseek, groq, xai, azure, cohere, claude_code)  --model, -m <model>           # Specific model to use  --web-search, -w              # Search web for API documentation  --auth <vars...>              # Environment variables for auth  --health-check                # Include health check endpoint (default: true)  --logging                     # Enable structured logging (default: true)  --metrics                     # Enable Prometheus metrics  --rate-limit <n>              # Rate limiting (requests per minute)  --retries                     # Enable retry logic (default: true)  --budget <amount>             # Maximum spend in USD (aborts if exceeded)  --compare-costs               # Show cost comparison across providers before generating

Configuration

Environment Variables

LLM Providers

All providers use the Vercel AI SDK via a unified UnifiedLLMProvider class with lazy dynamic imports — only the @ai-sdk/* package for your chosen provider is loaded at runtime.

Programmatic Usage

Basic Usage

import { ToolFactoryAgent, writeServerToDirectory, formatCost } from '@heshamfsalama/mcp-tool-factory';
// Create agent (auto-detects provider from env vars)const agent = new ToolFactoryAgent();
// Generate from descriptionconst server = await agent.generateFromDescription(  'Create tools for managing a todo list with priorities',  {    serverName: 'todo-server',    webSearch: true,    parallel: true,           // Enable parallel generation (default)    maxConcurrency: 5,        // Max concurrent LLM calls (default)    budget: 1.00,             // Optional: abort if cost exceeds $1.00    productionConfig: {      enableLogging: true,      enableMetrics: true,    },  });
// Cost tracking — see how much the generation costif (server.executionLog) {  console.log(`Cost: ${formatCost(server.executionLog.totalCost)}`);}
// Write to directoryawait writeServerToDirectory(server, './servers/todo');

From OpenAPI

import { ToolFactoryAgent, writeServerToDirectory } from '@heshamfsalama/mcp-tool-factory';import { readFileSync } from 'fs';import yaml from 'js-yaml';
const spec = yaml.load(readFileSync('./openapi.yaml', 'utf-8'));const agent = new ToolFactoryAgent({ requireLlm: false });
const server = await agent.generateFromOpenAPI(spec, {  serverName: 'my-api-server',  baseUrl: 'https://api.example.com',});
await writeServerToDirectory(server, './servers/api');

From Database

import { ToolFactoryAgent, writeServerToDirectory } from '@heshamfsalama/mcp-tool-factory';
const agent = new ToolFactoryAgent({ requireLlm: false });
// SQLite (auto-detected from file path)const server = await agent.generateFromDatabase('./data/app.db', {  serverName: 'app-database-server',  tables: ['users', 'posts', 'comments'],});
// PostgreSQL (auto-detected from connection string)const pgServer = await agent.generateFromDatabase(  'postgresql://user:pass@localhost/mydb',  { serverName: 'postgres-server' });
await writeServerToDirectory(server, './servers/app-db');

From GraphQL

import { ToolFactoryAgent, writeServerToDirectory } from '@heshamfsalama/mcp-tool-factory';import { readFileSync } from 'fs';
const schema = readFileSync('./schema.graphql', 'utf-8');const agent = new ToolFactoryAgent({ requireLlm: false });
const server = await agent.generateFromGraphQL(schema, {  serverName: 'my-graphql-server',});
await writeServerToDirectory(server, './servers/graphql');

From Ontology

import { ToolFactoryAgent, writeServerToDirectory } from '@heshamfsalama/mcp-tool-factory';import { readFileSync } from 'fs';
const ontologyData = readFileSync('./ontology.owl', 'utf-8');const agent = new ToolFactoryAgent({ requireLlm: false });
const server = await agent.generateFromOntology(ontologyData, {  serverName: 'knowledge-server',  format: 'rdf',});
await writeServerToDirectory(server, './servers/knowledge');

Code Validation

import { validateTypeScriptCode, validateGeneratedServer } from '@heshamfsalama/mcp-tool-factory';
// Validate TypeScript syntaxconst result = await validateTypeScriptCode(code);// { valid: false, errors: [{ line: 4, column: 1, message: "'}' expected." }] }
// Validate complete serverconst serverResult = await validateGeneratedServer(serverCode);// { valid: true, errors: [], summary: 'Generated server code is syntactically valid' }

Use with AI Frameworks

Claude Code / Claude Desktop

Add to your MCP settings (claude_desktop_config.json):

{  "mcpServers": {    "my-server": {      "command": "npx",      "args": ["tsx", "./servers/my-server/src/index.ts"]    }  }}

OpenAI Agents SDK

from agents import Agentfrom agents.mcp import MCPServerStdio
async with MCPServerStdio(    command="npx",    args=["tsx", "./servers/my-server/src/index.ts"]) as mcp:    agent = Agent(        name="My Agent",        tools=mcp.list_tools()    )

Google ADK

from google.adk.tools.mcp_tool import MCPToolset
tools = MCPToolset(    connection_params=StdioServerParameters(        command="npx",        args=["tsx", "./servers/my-server/src/index.ts"]    ))

LangChain

from langchain_mcp_adapters.client import MCPClient
client = MCPClient(    command="npx",    args=["tsx", "./servers/my-server/src/index.ts"])tools = client.get_tools()

Production Features

Structured Logging

mcp-factory generate "..." --logging

Generates servers with pino structured JSON logging:

const logger = pino({ level: 'info' });logger.info({ tool: 'get_weather', params }, 'Tool called');

Prometheus Metrics

mcp-factory generate "..." --metrics

Generates servers with prom-client metrics:

• mcp_tool_calls_total - Counter of tool invocations
• mcp_tool_duration_seconds - Histogram of execution times

Rate Limiting

mcp-factory generate "..." --rate-limit 100

Configurable rate limiting per client with sliding window.

Retry Logic

mcp-factory generate "..." --retries

Exponential backoff retry for transient failures.

Structured Error Codes

Generated servers use structured error codes for consistent error handling:

• INVALID_INPUT - Malformed or invalid tool parameters
• NOT_FOUND - Requested resource does not exist
• AUTH_ERROR - Authentication or authorization failure
• INTERNAL_ERROR - Unexpected server error

Enhanced Health Check

The /health endpoint returns detailed server status:

{  "status": "ok",  "version": "1.0.0",  "uptime": 3600,  "memory": { "rss": 52428800, "heapUsed": 20971520 },  "transport": "streamable-http"}

MCP Registry Publishing

Publish your generated servers to the MCP Registry for discoverability.

Generate with Registry Support

mcp-factory generate "Create weather tools" \  --name weather-server \  --github-username your-github-username \  --description "Weather tools for Claude" \  --version 1.0.0

This generates registry-compliant files:

package.json:

{  "name": "@your-github-username/weather-server",  "mcpName": "io.github.your-github-username/weather-server"}

server.json:

{  "$schema": "https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json",  "name": "io.github.your-github-username/weather-server",  "packages": [{    "registryType": "npm",    "identifier": "@your-github-username/weather-server",    "transport": { "type": "stdio" }  }],  "tools": [...]}

Publish Workflow

# 1. Build and publish to npmcd ./servers/weather-servernpm install && npm run buildnpm publish --access public
# 2. Install mcp-publisherbrew install modelcontextprotocol/tap/mcp-publisher
# 3. Authenticatemcp-publisher login github
# 4. Publish to registrymcp-publisher publish

See Publishing Guide for detailed instructions.

Architecture

┌───────────────────────────────────────────────────────────────────────┐│                         MCP Tool Factory                               │├───────────────────────────────────────────────────────────────────────┤│  Input Sources                                                         ││  ┌───────────┐ ┌───────────┐ ┌───────────┐ ┌───────────┐ ┌─────────┐││  │ Natural   │ │  OpenAPI  │ │ Database  │ │ GraphQL   │ │Ontology │││  │ Language  │ │   Spec    │ │  Schema   │ │  Schema   │ │RDF/YAML │││  └─────┬─────┘ └─────┬─────┘ └─────┬─────┘ └─────┬─────┘ └────┬────┘││        └──────────┬───┴─────────────┴─────────────┴────────────┘     ││                   ▼                                                    ││  ┌────────────────────────────────────────────────────────────────┐   ││  │                     ToolFactoryAgent                            │   ││  │  ┌─────────────────────────────────────────────────────────┐   │   ││  │  │  UnifiedLLMProvider (Vercel AI SDK)                      │   │   ││  │  │  Anthropic │ OpenAI │ Google │ Mistral │ DeepSeek       │   │   ││  │  │  Groq │ xAI │ Azure │ Cohere + Claude Code OAuth       │   │   ││  │  └─────────────────────────────────────────────────────────┘   │   ││  │  ┌──────────────┐ ┌──────────────┐ ┌───────────────────────┐  │   ││  │  │  LLM Cache   │ │  Cost        │ │ Parallel Generation   │  │   ││  │  │  (TTL-based) │ │  Tracking    │ │ (max concurrency: 5)  │  │   ││  │  └──────────────┘ └──────────────┘ └───────────────────────┘  │   ││  └────────────────────────────────────────────────────────────────┘   ││                   │                                                    ││                   ▼                                                    ││  ┌────────────────────────────────────────────────────────────────┐   ││  │                       Generators                                │   ││  │  ServerGenerator  │  DocsGenerator  │  TestsGenerator          │   ││  └────────────────────────────────────────────────────────────────┘   ││                   │                                                    ││                   ▼                                                    ││  ┌────────────────────────────────────────────────────────────────┐   ││  │                     GeneratedServer                             │   ││  │  Tools │ Resources │ Prompts │ Tests │ Docs │ Dockerfile       │   ││  └────────────────────────────────────────────────────────────────┘   ││                   │                                                    ││                   ▼                                                    ││  ┌────────────────────────────────────────────────────────────────┐   ││  │                Streamable HTTP Transport                        │   ││  │          POST /mcp  │  GET /health                             │   ││  └────────────────────────────────────────────────────────────────┘   │└───────────────────────────────────────────────────────────────────────┘

Development

# Clone the repositorygit clone https://github.com/HeshamFS/mcp-tool-factory-ts.gitcd mcp-tool-factory-ts
# Install dependenciespnpm install
# Buildpnpm run build
# Run testspnpm test
# Type checkpnpm run typecheck
# Lintpnpm run lint

Project Structure

mcp-tool-factory-ts/├── src/│   ├── agent/              # Main ToolFactoryAgent│   ├── auth/               # OAuth2 providers│   ├── cache/              # LLM response caching with configurable TTL│   ├── cli/                # Command-line interface│   ├── config/             # Configuration management│   ├── database/           # Database introspection (SQLite, PostgreSQL)│   ├── execution-logger/   # Execution logging│   ├── generators/         # Code generators (server, docs, tests)│   ├── graphql/            # GraphQL SDL parsing and server generation│   ├── middleware/         # Validation middleware│   ├── models/             # Data models│   ├── observability/      # Telemetry and tracing│   ├── ontology/           # Ontology parsing (RDF/OWL, JSON-LD, YAML)│   ├── openapi/            # OpenAPI spec parsing│   ├── production/         # Production code generation│   ├── prompts/            # LLM prompt templates│   ├── providers/          # LLM providers (10 providers via Vercel AI SDK + Claude Code)│   ├── security/           # Security scanning│   ├── server/             # MCP server mode (factory-as-a-server)│   ├── templates/          # Handlebars templates for generated files│   ├── validation/         # Code validation and Zod schemas│   └── web-search/         # Web search integration├── docs/                   # Documentation├── tests/                  # Test files└── dist/                   # Built output

Documentation

• Getting Started
• CLI Reference
• API Reference
• Examples
• OpenAPI Guide
• Database Guide
• Providers Guide
• Production Features
• Architecture
• Troubleshooting
• Contributing

Troubleshooting

Common Issues

API Key Not Found

# Check your environmentecho $ANTHROPIC_API_KEY
# Set itexport ANTHROPIC_API_KEY=your-key-here

Generated Server Won't Start

# Install dependencies firstcd ./servers/my-servernpm installnpx tsx src/index.ts

TypeScript Errors

# Validate generated codeimport { validateGeneratedServer } from '@heshamfsalama/mcp-tool-factory';const result = await validateGeneratedServer(code);console.log(result.errors);

See Troubleshooting Guide for more solutions.

Changelog

v0.3.0

• Vercel AI SDK Migration - All LLM providers now use the Vercel AI SDK via a single UnifiedLLMProvider class with lazy dynamic imports. Removed ~473 LOC of provider-specific implementations. Only the @ai-sdk/* package for your chosen provider is loaded at runtime.
• 10 LLM Providers - Added Mistral, DeepSeek, Groq, xAI, Azure, and Cohere alongside existing Anthropic, OpenAI, Google, and Claude Code providers. All use the same unified interface.
• Cost Tracking - Every LLM call now calculates estimated cost using a built-in pricing table for 50+ models. Shows per-call cost, total generation cost, and per-phase breakdown (tool extraction, implementation, tests, docs). Detailed token breakdowns include cache read/write tokens and reasoning tokens from the AI SDK.
• Budget Limits (--budget <amount>) - Set a maximum spend in USD. Generation aborts gracefully with BudgetExceededError if cumulative cost exceeds the budget.
• Provider Cost Comparison (--compare-costs) - Before generation, estimates cost across all available providers and shows a sorted comparison table. No extra API calls needed — uses the static pricing table.
• Per-Phase Cost Breakdown - CLI output and execution logs show which generation steps cost the most (tool extraction, implementation, resource extraction, prompt extraction, test generation, docs generation).
• OpenAI Reasoning Model Support - Temperature parameter is automatically omitted for OpenAI o-series and gpt-5.x models that don't support it.

v0.2.0

• Streamable HTTP Transport - Generated servers use StreamableHTTPServerTransport with native http module instead of Express/SSE (deprecated June 2025). Single /mcp POST endpoint with /health GET endpoint.
• MCP SDK v1.26.0 - Updated from ^1.0.0 to ^1.26.0
• Resources & Prompts - Full support for all three MCP primitives. Resources expose structured data (documents, DB records, file trees). Prompts provide reusable templates for guided LLM workflows. Agent automatically extracts resources and prompts from descriptions via LLM.
• GraphQL Input Source - New from-graphql CLI command and generate_from_graphql MCP tool. Queries map to read tools, mutations map to write tools, and GraphQL types are converted to Zod schemas.
• Ontology Input Source - New from-ontology CLI command and generate_from_ontology MCP tool. Supports RDF/OWL, JSON-LD, and custom YAML formats. OWL Classes map to Resources, ObjectProperties to Tools, DataProperties to tool parameters.
• LLM Response Caching - Deduplicates identical LLM calls with configurable TTL. Bypass with skipCache option.
• Parallel Generation - Tool implementations generated concurrently by default (parallel: true, maxConcurrency: 5). Significant speed improvement for multi-tool servers.
• InMemoryTransport Testing - Generated tests use InMemoryTransport.createLinkedPair() instead of subprocess spawning. Servers export createServer() factory function for testability.
• Production Enhancements - Rate limiting, structured logging, metrics, and duration tracking wired into tool handlers. Enhanced health check with version, uptime, memory, and transport info. Structured error codes: INVALID_INPUT, NOT_FOUND, AUTH_ERROR, INTERNAL_ERROR.

v0.1.0

• Initial TypeScript release
• Natural language generation with Claude, Claude Code, OpenAI, Google Gemini
• OpenAPI 3.0+ specification import
• Database CRUD generation (SQLite, PostgreSQL)
• Production features (logging, metrics, rate limiting)
• MCP Registry server.json generation
• TypeScript syntax validation
• Web search for API documentation
• GitHub Actions CI/CD generation
• MCP Server mode for on-the-fly generation with Claude

License

MIT

Links

• GitHub Repository
• npm Package
• MCP Specification
• MCP Registry
• Python Version