Model Context Protocol

AIHumanly MCP Service

Integrate AIHumanly's Humanizer directly into your AI coding assistants. Codex, Claude Code, TRAE, Cursor — one URL connects them all. Your AI tools can now humanize text like a local function call.

Overview

AIHumanly MCP (Model Context Protocol) service allows you to integrate AIHumanly's Humanizer capabilities directly into AI coding assistants (such as Codex, Claude Code, TRAE, Cursor, and more). AI coding assistants can call humanization as easily as a local tool — just like a universal USB interface for AI tools. Any MCP-compatible AI tool only needs one URL to connect.

Quick Start

1. Get Your API Key

Visit aihumanly.io, register or log in, then get your API Key (160-character hex string) from your dashboard. Make sure your account has an active subscription with available word credits.

2. Configure Your AI Tool

Choose your AI tool below and add the MCP configuration. Each tool has its own configuration format — pick yours and follow the guide.

Configuration Guides

Select your AI tool below to see the configuration guide.

Available Tools

humanize_text

Convert AI-generated text into more natural, human-like writing, reducing AI detection rates.

Parameters

NameTypeRequiredDescription
paragraphstringYesThe original paragraph text to convert (minimum 15 words).
langstringYesLanguage code (e.g., en, zh-cn, es, fr, de, jp).

Usage Example

In your AI coding assistant, just describe what you want in natural language: "Help me remove AI traces from this text: Artificial intelligence has revolutionized the way we interact with technology." The assistant will automatically call the humanize_text tool and return the humanized result.

Technical Architecture

How MCP connects your AI coding tools to AIHumanly's Humanizer engine.

AI Coding Assistant

Codex, Claude Code, TRAE, Cursor — your AI tool sends natural language requests.

MCP Protocol

JSON-RPC 2.0 over SSE — the standardized communication layer.

MCP Server (Bridge)

api.aihumanly.io/mcp — handles SSE connections, protocol translation, and authentication.

AIHumanly Backend

The core Humanizer engine — paragraph conversion, rewriting, vocabulary replacement, AI detection, similarity scoring.

MCP Endpoints

GET/mcp/sse

SSE connection endpoint — establishes persistent event stream.

POST/mcp/messages?sessionId=xxx

JSON-RPC message endpoint — sends requests and receives SSE responses.

POST/mcp

Streamable HTTP endpoint — compatible with newer MCP clients.

Supported MCP Methods

initialize

Handshake negotiation — exchanges client/server capabilities.

notifications/initialized

Client initialization complete notification.

tools/list

Retrieve available tools list.

tools/call

Call a specific tool.

ping

Heartbeat check.

Authentication

The MCP service supports two ways to pass your API Key:

Method 1: HTTP Header (Recommended)

Set the x-api-key header in your MCP config. Also supports standard Bearer Token format.

json
"headers": {
  "x-api-key": "YOUR_API_KEY"
}

Method 2: Tool Parameter

Pass api_key as a parameter when calling humanize_text. If both methods are used, the Header takes priority.

json
{
  "name": "humanize_text",
  "arguments": {
    "paragraph": "Your text here...",
    "lang": "en",
    "api_key": "YOUR_API_KEY"
  }
}

Limits & Quotas

  • •Each call consumes your subscription word balance (shared with the web app).
  • •Paragraphs must have a minimum of 15 words.
  • •Single request timeout: 180 seconds.
  • •API Key must be linked to an active subscription.
  • •If word balance is insufficient, the tool call returns an error.

Frequently Asked Questions

Common questions about the AIHumanly MCP Service.

Ready to Supercharge Your AI Coding Tools?

Connect AIHumanly to your favorite AI assistant in minutes. One URL, endless humanization.

Get Started Free