Base64 MCP Server
# Base64编码解码MCP服务器
[](https://smithery.ai/server/@liuyazui/base64_server)
[English Version](README_EN.md)
一个简单高效的MCP服务器,专注于提供Base64编码和解码功能,支持文本和图片的Base64转换。
<a href="https://glama.ai/mcp/servers/@liuyazui/base64_server">
<img width="380" height="200" src="https://glama.ai/mcp/servers/@liuyazui/base64_server/badge" alt="Base64 Server MCP server" />
</a>
## 功能特点
- 文本Base64编码和解码
- 图片Base64编码和解码
- 支持Data URL格式
- 简单易用的API
- 使用uv进行依赖管理
## 安装
### 使用uv安装
```bash
# 创建虚拟环境
uv venv
# 激活虚拟环境(Linux/macOS)
source .venv/bin/activate
# 激活虚拟环境(Windows)
.venv\Scripts\activate
# 安装包(开发模式)
uv pip install -e .
# 安装带开发依赖的包
uv pip install -e ".[dev]"
```
### 安装Smithery
使用Smithery为Claude桌面安装Base64编码解码MCP服务器,使用以下命令:
```bash
npx -y @smithery/cli install @liuyazui/base64_server --client claude
```
## 使用方法
### 使用MCP Inspector测试
```bash
# 使用MCP Inspector测试服务器
uv run mcp dev base64_server.py
```
### 与MCP client集成
1. 添加服务器配置:
```json
{
"mcpServers": {
"base64-encoder": {
"command": "uv",
"args": [
"run",
"--with",
"mcp[cli]",
"mcp",
"run",
"[path to base64_server.py]"
]
}
}
}
```
## API参考
### 工具(Tools)
- **base64_encode_text(text: str) -> str**:将文本转换为Base64编码
- **base64_decode_text(encoded: str) -> str**:将Base64编码解码为文本
- **base64_encode_image(image_path: str) -> str**:将图片转换为Base64编码
- **base64_decode_image(encoded: str, output_path: str, mime_type: str = "image/png") -> str**:将Base64编码解码为图片
### 资源(Resources)
- **encode://base64/text/{text}**:获取文本的Base64编码
- **decode://base64/text/{encoded}**:获取Base64编码的解码结果
- **encode://base64/image/{image_path}**:获取图片的Base64编码
- **decode://base64/image/{encoded}**:获取Base64编码的解码图片
### 提示模板(Prompts)
- **base64_usage_guide()**: 提供Base64服务的基本使用指南
- **encode_text_prompt(text: str)**: 文本编码提示模板
- **encode_image_prompt(image_path: str)**: 图片编码提示模板
- **error_handling_prompt(error_message: str)**: 错误处理提示模板
使用示例:
```python
# 获取使用指南提示
messages = await client.get_prompt("base64_usage_guide")
# 获取文本编码提示
messages = await client.get_prompt("encode_text_prompt", {"text": "Hello World"})
```
## 开发
## 许可证
MIT
TDQS
Scored across 4 tools
Every tool has a clearly distinct purpose with no ambiguity. The four tools cleanly separate into two decode operations (image vs text) and two encode operations (image vs text), each targeting specific data types and use cases. An agent can easily distinguish between them based on the input/output type and operation direction.
All tools follow a perfectly consistent verb_noun pattern with 'base64_' prefix, then operation (encode/decode), then target type (image/text). The naming is predictable, readable, and follows the same convention throughout without any deviations or mixed styles.
Four tools is ideal for this server's purpose. It provides complete coverage of the Base64 domain with exactly the right granularity: encode and decode operations for both text and image data types. No tool feels redundant or missing, and the count is well-scoped for the functionality offered.
The tool surface is complete for Base64 operations. It covers both encoding and decoding for the two primary data types (text and images) that Base64 typically handles. There are no gaps in the CRUD/lifecycle for this domain, and agents can perform all expected Base64 transformations without workarounds.