MCP Jina

English | 中文

English Documentation

Introduction

A Jina AI service implementation based on Model Context Protocol (MCP). This service enables LLMs like Claude to access Jina AI capabilities through the MCP protocol.

Features

• Full MCP protocol support
• Jina AI API integration
• Tool invocation capabilities
• Claude Desktop integration

Requirements

• Python >= 3.12
• Valid Jina AI API key

Quick Start

1. Installation

# Clone repository
git clone https://github.com/youqingkui/mcp_jina.git
cd mcp_jina

# Create environment and install dependencies
uv venv
source .venv/bin/activate  # Linux/Mac
# or .venv\Scripts\activate  # Windows

uv pip install -e .

2. Configuration

# Copy environment template
cp .env.example .env

# Edit .env file, add your API key
JINA_API_KEY=your_jina_api_key_here

3. Test Service

Use MCP Inspector for testing:

npx @modelcontextprotocol/inspector uv run mcp-jina

Claude Desktop Integration

1. Configuration File Location

• MacOS/Linux: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

2. Configuration Example

{
    "mcpServers": {
        "mcp-jina": {
            "command": "uv",
            "args": [
                "--directory",
                "/absolute/path/to/mcp_jina",  // Replace with actual path
                "run",
                "mcp-jina"
            ],
            "env": {
                "JINA_API_KEY": "your_jina_api_key_here"
            }
        }
    }
}

3. Usage

1. Launch Claude Desktop
2. Check the hammer icon in the toolbar for service connection status
3. Use Jina AI features in conversations

Troubleshooting

Common Issues

• Service Not Showing
  • Check configuration file syntax
  • Verify project path
  • Restart Claude Desktop

View Logs

# MacOS/Linux
tail -f logs/jina_reader.log

# Windows
type logs\jina_reader.log

Common Error Solutions

• Invalid API Key: Check key in .env file
• Connection Failed: Check network and firewall settings
• Service Start Failed: Verify Python environment setup

---

中文文档

简介

基于 Model Context Protocol (MCP) 的 Jina AI 服务实现。该服务允许 Claude 等 LLM 通过 MCP 协议访问 Jina AI 的功能。

功能特点

• 完整支持 MCP 协议标准
• 集成 Jina AI API 功能
• 提供工具调用能力
• 支持 Claude Desktop 集成

系统要求

• Python >= 3.12
• 有效的 Jina AI API 密钥

快速开始

1. 安装

# 克隆仓库
git clone https://github.com/youqingkui/mcp_jina.git
cd mcp_jina

# 创建环境并安装依赖
uv venv
source .venv/bin/activate  # Linux/Mac
# 或 .venv\Scripts\activate  # Windows

uv pip install -e .

2. 配置

# 复制环境变量模板
cp .env.example .env

# 编辑 .env 文件，添加你的 API 密钥
JINA_API_KEY=your_jina_api_key_here

3. 测试服务

使用 MCP Inspector 进行测试：

npx @modelcontextprotocol/inspector uv run mcp-jina

Claude Desktop 集成

1. 配置文件位置

• MacOS/Linux: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

2. 配置示例

{
    "mcpServers": {
        "mcp-jina": {
            "command": "uv",
            "args": [
                "--directory",
                "/absolute/path/to/mcp_jina",  // 替换为实际路径
                "run",
                "mcp-jina"
            ],
            "env": {
                "JINA_API_KEY": "your_jina_api_key_here"
            }
        }
    }
}

3. 使用方法

1. 启动 Claude Desktop
2. 检查工具栏中的锤子图标，确认服务连接状态
3. 在对话中使用 Jina AI 功能

故障排除

常见问题

• 服务未显示
  • 检查配置文件语法
  • 确认项目路径正确
  • 重启 Claude Desktop

日志查看

# MacOS/Linux
tail -f logs/jina_reader.log

# Windows
type logs\jina_reader.log

常见错误解决

• API 密钥无效：检查 .env 文件中的密钥
• 连接失败：检查网络和防火墙设置
• 服务启动失败：验证 Python 环境配置