Build hosted MCP serversfor real APIs, from the UIor your AI agent

apitomcp.ai · 2026-06-20 · by Long Nguyen ·generic
Build hosted remote HTTP MCP servers from REST and GraphQL APIs with OAuth, API keys, workflow tools, encrypted credentials, snapshots, and an AI agent builder.

Build hosted MCP serversfor real APIs, from the UIor your AI agent

Hosted MCP servers for real APIs

Build hosted MCP serversfor real APIs, from the UIor your AI agent

Turn REST and GraphQL APIs from public services, SaaS platforms, and internal business systems into hosted remote HTTP MCP servers with OAuth, API key, Bearer token, Basic Auth, workflow tools, encrypted credentials, and JMESPath output mapping.

Start with Visual BuilderSet Up Agent Builder

Enterprise Security

UI or Agent Builder

Deploy in Minutes

Compatible with

ChatGPT

ChatGPT

Claude

Claude

Codex

Codex

Cursor

Cursor

Claude Code

Claude Code

Visual Studio Code

VS Code

Antigravity

Antigravity

Custom Agents

Custom Agents

From API Endpoint

GET

https://api.company.com/revenue

{ "revenue": { "monthly": 125000, "yearly": 1500000, "growth_rate": 15.2, "currency": "USD" } }

To MCP Server

Tools: get\_revenue

Created by UI or your connected AI agent

Remote HTTP URL ready for Codex, Cursor, Claude Code

AI Agent Integration

QUESTION

"What's our current revenue performance?"

AI Agent: "Monthly revenue is $125,000, yearly revenue is $1.5M, and growth rate is 15.2%."

Visual + Agent Builder

Build Path

OAuth + API Keys

Authentication

Workflow Tools

Tooling

REST + GraphQL

API Types

Build visually, or build from your agent

Two production-ready ways to create MCP servers

Use the dashboard when you want hands-on control. Use Agent Builder when you want Codex, Cursor, Claude Code, or another IDE agent to create and iterate for you.

Visual Builder

A guided dashboard for teams that want to review auth, tools, output mapping, test results, and deployment settings before publishing.

  • Configure auth, base URL, and MCP access
  • Define API tools, workflow tools, and output mapping
  • Test requests before deployment
  • Deploy a production MCP endpoint

    • Dashboard flow

    API

    Tools

    Test

    Deploy

    Start with Visual Builder

    AI Agent Builder

    Connect the API To MCP manager server once, then ask your coding agent to create, update, test, deploy, and inspect servers from chat.

  • Connect https://mcp.apitomcp.io/ to your agent
  • Create a scoped manager token
  • Describe the API you want to convert
  • Let the agent create, update, test, and deploy it

    • Manager MCP URL

    `https://mcp.apitomcp.io/`

    Prompt example

    Create an MCP server for our internal support platform using OAuth Authorization Code so each employee connects their own account. Add ticket lookup and workflow tools, test them, then return the MCP URL.

    Set Up Agent Builder

    What you can build

    Start from your own business platform, internal endpoints, authenticated SaaS APIs, or public data APIs. API To MCP turns them into hosted tools your agents can call.

    Business Platforms

    Expose your company CRM, ERP, HR, finance, support, or operations platform to employees through controlled MCP tools.

    Internal CRMERPHRISFinance Ops

    Marketing and SEO

    Connect ad, analytics, and search APIs for reporting, inspection, content, and optimization workflows.

    Meta AdsGoogle AdsGoogle AnalyticsSearch Console

    Commerce and Billing

    Build MCP tools around billing, catalog, product, and order APIs.

    PayPalShopifyWooCommerce

    Developer

    Give coding agents controlled access to repositories, deploys, issues, and observability data.

    GitHubGitLabVercelSentry

    Public Data

    Publish no-auth MCP servers for open data APIs that anyone can use by URL.

    WeatherREST CountriesWorld BankHacker News

    Content

    Turn content systems into MCP tools for publishing, lookup, and editorial workflows.

    WordPress.comContentfulWebflowNotion

    Authentication for real-world APIs

    API To MCP separates upstream API authentication from MCP server access, so teams can support REST and GraphQL APIs, private systems, and employee-authorized platforms.

    Credential security

    Stored API keys, Bearer tokens, Basic passwords, OAuth client secrets, access tokens, and refresh tokens are encrypted at rest and masked in the owner UI.

    Encrypted at rest for stored credentials and OAuth tokens

    Masked in edit screens unless the owner replaces the value

    Snapshots never include live secrets or active connection tokens

    No Authentication

    For public APIs and open-data endpoints that do not require credentials.

    API Key

    Send keys in headers or query parameters for services that use token-style access.

    Bearer Token

    Store a token securely, or forward the Bearer token supplied by the connected agent.

    Basic Auth

    Use username and password authentication for legacy or internal APIs.

    OAuth Client Credentials

    Use machine-to-machine OAuth for billing, commerce, operations, and internal services.

    OAuth Authorization Code

    Let every employee or end user connect their own upstream account without an API To MCP account.

    MCP access modes

    Choose how AI clients connect to the MCP server itself. This is separate from the credential used to call the upstream API.

    Open MCP servers for public or agent-supplied upstream credentials

    OAuth/Bearer Token for clients that authenticate the MCP connection

    Client Token for an extra API To MCP access layer when needed

    Everything you need to convert APIs to MCP

    Build, authenticate, compose, deploy, and share production MCP servers without writing custom MCP runtime code.

    Visual MCP Builder

    Configure REST or GraphQL integrations, tools, schemas, response mapping, tests, and deployment from a guided dashboard.

    Manager MCP for AI Agents

    Connect the API To MCP manager server and let AI agents create, update, test, deploy, and inspect MCP servers from chat.

    OAuth Client Credentials

    Use server-to-server OAuth for APIs such as billing, commerce, internal services, and GraphQL backends that issue machine tokens.

    Per-user OAuth Authorization Code

    Let each employee or MCP end user connect their own upstream business-platform account through OAuth.

    Workflow Tools

    Compose multiple API tool calls into one MCP tool for multi-step lookup, enrichment, and action flows.

    Response Mapping with JMESPath

    Shape nested upstream API responses into clean JSON outputs that are easier for agents to consume.

    Templates and Public Directory

    Publish live MCP servers to the directory or share reusable snapshots that copy tools and configuration without secrets.

    Hosted Streamable HTTP Runtime

    Deploy remote HTTP MCP endpoints with Open or Authenticated access, SSL, usage tracking, and hosted runtime management.

    Compatible Platforms

    ChatGPT

    ChatGPT

    Claude

    Claude

    Codex

    Codex

    Cursor

    Cursor

    Claude Code

    Claude Code

    Visual Studio Code

    VS Code

    Antigravity

    Antigravity

    Custom Agents

    Custom Agents

    Ready to transform your APIs? Start from the dashboard or your AI agent.

    How it works

    Build MCP servers through the visual UI or directly from the AI agent where you already work.

    Lane A: Visual Builder

    Best when you want full control in the dashboard.

    Configure API

    Create tools

    Run tests

    Deploy server

    Lane B: Agent Builder

    Best when you want your IDE agent to do the setup.

    Connect once

    https://mcp.apitomcp.io/

    Ask your agent to create, test, and deploy the MCP server.

    1

    Configure API access

    Enter your REST or GraphQL API base URL and choose the upstream auth model: no auth, API key, Bearer token, Basic Auth, OAuth Client Credentials, or OAuth Authorization Code.

  • Base URL setup
  • Upstream authentication
  • MCP access policy

    • Create MCP Server

    Basic Information

    Server Name \*

    Your Site MCP

    A descriptive name for your MCP server

    Description

    Describe what this MCP server does and what APIs it integrates with...

    Version

    1.0.0

    Server URL Preview

    `https://your-site-mcp.apitomcp.io`

    This will be your MCP server's production endpoint

    API Integration

    Base API URL \*

    https://api.yoursite.com/v1

    The root URL of the API you want to integrate

    Authentication Type

    No Authentication

    API Key

    Bearer Token

    Basic Auth

    OAuth Client Credentials / Authorization Code

    Bearer Token

    ••••••••••••••••••••••••••••••••••••••••

    Will be sent as: Authorization: Bearer <token>

    Custom Headers (Optional)Add Header

    2

    Define tools and output

    Create API tools or workflow tools, define input schemas, run endpoint tests, and shape responses with JMESPath when needed.

  • API and workflow tools
  • Parameter validation
  • JMESPath response mapping

    • Configure Tools

    Tools (1)

    get\_ordersActive

    Get list orders

    2 params

    Edit Tool 1

    Configure MCP tool according to specification

    Basic InfoInput SchemaAdvanced

    Input Parameters Schema\+ Add Parameter

    start\_dateRequired

    Type

    String

    Description

    Start Date (format: YYYY-MM-DD)

    end\_dateRequired

    Type

    String

    Description

    End Date (format: YYYY-MM-DD)

    JSON Schema Preview

    {

    "type": "object",

    "properties": {

    "start\_date": { "type": "string" },

    "end\_date": { "type": "string" }

    }

    }

    3

    Deploy & Connect

    Deploy your MCP server to the hosted Streamable HTTP runtime and connect it to remote MCP clients.

  • Hosted MCP endpoint
  • Open or authenticated access
  • AI agent integration

Integration Guide

Integration Guide

Your MCP server is now live! Use the configuration below to connect it to AI clients:

Remote MCP Configuration

VS Code / Cursor

{

"servers": {

"your-site-mcp": {

"type": "http",

"url": "https://your-site-mcp-w8jg832k998m.apitomcp.io/"

}

}

}

Setup Instructions

For remote HTTP MCP clients:

1. Copy your server URL 2. Add it as a remote HTTP MCP server 3. Add a Bearer token only when the server access policy requires one 4. Enable the server in your AI agent or IDE 5. Your tools are available to the agent immediately

Your MCP Server URL:

`https://your-site-mcp-w8jg832k998m.apitomcp.io/`

This URL will be available after you deploy your MCP server.

Start Building Your MCP Server

Build from the dashboard or directly from your AI agent

Ready to transform yourAPIs into MCP servers?

Start building powerful AI integrations today. Connect any REST or GraphQL API to ChatGPT, Claude, Codex, Cursor, Claude Code, and other MCP clients in minutes.

Open Visual BuilderSet Up Agent Builder

Fast setup

Create your first MCP server from a guided flow

Access controls

Use Open, OAuth/Bearer Token, or Client Token MCP access

Hosted runtime

Serve MCP tools from a managed remote HTTP URL

No credit card required • Free tier available • Cancel anytime

SSL endpoints

Usage tracking

Free plan available

apitomcp.ai · 15 links

Comments (0)

No comments yet