- Explore MCP Servers
- doubao_mcp_server
Doubao Mcp Server
What is Doubao Mcp Server
Doubao MCP Server is a Model Context Protocol (MCP) server based on the Doubao API from Volcano Engine, designed to provide powerful multimodal generation capabilities for AI clients. It integrates advanced AI features for generating images, videos, and more, enabling users to leverage these capabilities within their preferred development environments.
Use cases
The primary use cases include generating high-quality images from text descriptions, creating videos from text inputs, and producing dynamic videos from images combined with text prompts. This tool is beneficial for content creators, developers, and businesses looking to enhance their visual content effortlessly using AI.
How to use
To use Doubao MCP Server, install it via pip or uvx, and configure your AI client (such as Cursor, Cherry Studio, or Claude Desktop) to point to the MCP server. Set your Doubao API key in the configuration file of the client. You can then use various tools like ‘text_to_image’, ‘text_to_video’, ‘image_to_video’, and ‘encode_image_to_base64’ to generate content based on your inputs.
Key features
Key features include the capability to generate images and videos from text, create videos from images and text, convert local images to base64 encoding, and support for multiple AI models tailored for specific tasks. The server is also configurable, allowing users to select models based on their requirements.
Where to use
Doubao MCP Server can be used in various applications including creative design, marketing material creation, multimedia content development, and any domain that requires the generation of high-quality visual assets from text or image inputs. Its integration with different AI clients makes it accessible in diverse environments.
Overview
What is Doubao Mcp Server
Doubao MCP Server is a Model Context Protocol (MCP) server based on the Doubao API from Volcano Engine, designed to provide powerful multimodal generation capabilities for AI clients. It integrates advanced AI features for generating images, videos, and more, enabling users to leverage these capabilities within their preferred development environments.
Use cases
The primary use cases include generating high-quality images from text descriptions, creating videos from text inputs, and producing dynamic videos from images combined with text prompts. This tool is beneficial for content creators, developers, and businesses looking to enhance their visual content effortlessly using AI.
How to use
To use Doubao MCP Server, install it via pip or uvx, and configure your AI client (such as Cursor, Cherry Studio, or Claude Desktop) to point to the MCP server. Set your Doubao API key in the configuration file of the client. You can then use various tools like ‘text_to_image’, ‘text_to_video’, ‘image_to_video’, and ‘encode_image_to_base64’ to generate content based on your inputs.
Key features
Key features include the capability to generate images and videos from text, create videos from images and text, convert local images to base64 encoding, and support for multiple AI models tailored for specific tasks. The server is also configurable, allowing users to select models based on their requirements.
Where to use
Doubao MCP Server can be used in various applications including creative design, marketing material creation, multimedia content development, and any domain that requires the generation of high-quality visual assets from text or image inputs. Its integration with different AI clients makes it accessible in diverse environments.
Content
Doubao MCP Server
项目简介
Doubao MCP Server 是一个基于火山引擎豆包API的MCP(Model Context Protocol)服务器,为AI客户端提供强大的多模态生成能力。该项目的开发初衷是为了将火山引擎豆包的先进AI生成能力集成到支持MCP协议的各种AI客户端中,让用户能够在熟悉的开发环境中直接使用文生图、文生视频、图生视频等功能。
核心功能特性:
- 文生图:根据文本描述生成高质量图片
- 文生视频:根据文本描述生成视频内容
- 图生视频:基于图片和文本描述生成动态视频
- 图片编码:支持本地图片文件转换为base64编码
- 模型配置:支持多种豆包AI模型选择
支持的AI模型:
- 文生图模型:
doubao-seedream-3-0-t2i-250415
- 图生视频模型:
doubao-seedance-1-0-lite-i2v-250428
- 文生视频模型:
doubao-seedance-1-0-lite-t2v-250428
部署指南
环境依赖
- Python >= 3.13
- 火山引擎豆包API密钥
安装方式
方式一:使用pip安装
pip install doubao-mcp-server
方式二:使用uvx安装(推荐)
uvx doubao-mcp-server
客户端配置
Cursor配置
在 ~/.cursor/mcp.json
文件中添加以下配置:
{
"mcpServers": {
"doubao-mcp-server": {
"command": "uvx",
"args": [
"doubao-mcp-server"
],
"env": {
"DOUBAO_API_KEY": "your-api-key-here"
}
}
}
}
Cherry Studio配置
-
打开 Cherry Studio
-
进入 设置 → MCP Servers → 添加服务器
-
配置参数:
- 名称:
doubao-mcp-server
- 描述:
豆包AI生成服务
- 类型:
STDIO
- 命令:
uvx
- 参数:
doubao-mcp-server
- 环境变量:
DOUBAO_API_KEY=your-api-key-here
- 名称:
-
点击保存并启用
详细图解
配置好后,可以检查到工具有哪些
Claude Desktop配置
在 claude_desktop_config.json
文件中添加:
{
"mcpServers": {
"doubao-mcp-server": {
"command": "uvx",
"args": [
"doubao-mcp-server"
],
"env": {
"DOUBAO_API_KEY": "your-api-key-here"
}
}
}
}
Continue.dev配置
在 config.json
文件中添加:
{
"mcpServers": [
{
"name": "doubao-mcp-server",
"command": "uvx",
"args": [
"doubao-mcp-server"
],
"env": {
"DOUBAO_API_KEY": "your-api-key-here"
}
}
]
}
API密钥获取
-
访问 火山引擎控制台
-
注册并登录账户
-
开通豆包大模型服务(需要分别授权开通各个模型)
火山引擎模型,需要分别授权开通,点击开通管理
-
在API管理中创建API密钥
API 管理
可用工具
1. set_api_key
设置豆包API密钥
api_key
(string): 豆包API密钥
2. text_to_image
根据文本描述生成图片
prompt
(string): 图片描述提示词size
(string, 可选): 图片尺寸,默认"1024x1024"model
(string, 可选): 模型名称
支持的图片尺寸: 512x512, 768x768, 1024x1024, 1024x1792, 1792x1024
3. text_to_video
根据文本描述生成视频
prompt
(string): 视频描述提示词duration
(string, 可选): 视频时长(秒),默认"5"ratio
(string, 可选): 视频比例,默认"16:9"model
(string, 可选): 模型名称
支持的视频比例: 16:9, 9:16, 1:1
4. image_to_video
根据图片和文本描述生成视频
prompt
(string): 视频描述提示词image_base64
(string): 图片的base64编码字符串duration
(string, 可选): 视频时长(秒),默认"5"ratio
(string, 可选): 视频比例,默认"16:9"model
(string, 可选): 模型名称
5. encode_image_to_base64
将本地图片文件编码为base64字符串
image_path
(string): 图片文件路径
使用示例
文生图示例
请使用text_to_image工具生成一张"夕阳下的海边风景"图片
文生视频示例
请使用text_to_video工具生成一个"猫咪在花园里玩耍"的5秒视频
图生视频示例
首先使用encode_image_to_base64将图片编码,然后使用image_to_video生成视频
图解案例(文生生图)
文生视频
生成的视频
注意事项
- 视频生成任务可能需要较长时间完成,请耐心等待
- 确保API密钥有足够的配额
- 生成的内容URL有时效性,请及时保存
故障排除
常见问题
- API密钥错误: 确保API密钥正确且有效
- 网络连接问题: 检查网络连接和防火墙设置
- 模型不可用: 确认使用的模型名称正确
调试模式
启用详细日志输出:
uvx doubao-mcp-server --verbose
项目信息
许可证: MIT License
作者: wwzhouhui - [email protected]
版本: v0.1.0
- 初始版本发布
- 支持文生图、文生视频、图生视频功能
- 集成火山引擎豆包API
贡献: 欢迎提交Issue和Pull Request来改进这个项目