Vivado MCP Server
Lets LLM clients control a persistent AMD/Xilinx Vivado TCL session on Windows, avoiding repeated startup cost with native process handling.
README
Vivado MCP Server (Windows-Native)
Vivado MCP Server is a Model Context Protocol (MCP) server that lets LLM clients control a persistent AMD/Xilinx Vivado TCL session on Windows.
It is designed to avoid the pexpect dependency and run reliably on native
Windows by using subprocess.Popen plus a background stdout reader thread.
Why use this
- Persistent Vivado process to avoid repeated startup cost.
- Windows-native process handling (no POSIX pseudo-terminal requirement).
- Clean tool API for MCP clients.
- No command prompt popup windows (
CREATE_NO_WINDOW).
Requirements
- Windows 10/11
- Python 3.10+
- Vivado installed locally
uv(recommended) orpip
Quick install
Option A: local development install (uv)
git clone <your-repo-url>
cd vivado-mcp-win
uv sync
Option B: install from package index (when published)
pip install vivado-mcp-win
Vivado path configuration
The server resolves vivado.bat in this order:
start_sessiontool argumentvivado_path- Environment variable
VIVADO_BAT_PATH - PATH lookup (
vivado.batorvivado) - Fallback
C:\VIVADO\2025.2\Vivado\bin\vivado.bat
Recommended: set VIVADO_BAT_PATH explicitly.
PowerShell example:
$env:VIVADO_BAT_PATH = "C:\Xilinx\Vivado\2025.2\bin\vivado.bat"
Run the MCP server
From source checkout:
uv run vivado-mcp
From installed package:
vivado-mcp
Note: this is a stdio MCP server, so it waits for MCP messages from a client. Running it directly in a terminal will appear idle.
MCP client configuration
Use the vivado-mcp command as a stdio server in your MCP client config.
Generic example:
{
"mcpServers": {
"vivado": {
"command": "vivado-mcp",
"args": [],
"env": {
"VIVADO_BAT_PATH": "C:\\VIVADO\\2025.2\\Vivado\\bin\\vivado.bat"
}
}
}
}
If your client runs inside this repository, you can also use:
{
"mcpServers": {
"vivado": {
"command": "uv",
"args": ["run", "vivado-mcp"],
"env": {
"VIVADO_BAT_PATH": "C:\\VIVADO\\2025.2\\Vivado\\bin\\vivado.bat"
}
}
}
}
Exposed MCP tools
start_session: starts or reuses a persistent Vivado session.run_tcl_command: executes TCL in the live session.session_status: returns session status and metrics.stop_session: stops the running session.
Development
Install dev dependencies:
uv sync --group dev
Run lint:
uv run ruff check .
Run tests:
uv run pytest
Publish checklist
- Update project URLs in
pyproject.toml. - Bump version.
- Build package:
uv build
- Publish (example with trusted publisher or token):
uv publish
Troubleshooting
- Error
Vivado executable not found: setVIVADO_BAT_PATHor passvivado_pathtostart_session. - Session timeout on startup: increase startup timeout in code or ensure Vivado installation is healthy and licensed.
- Import errors in editor for
mcp.*: runuv syncand ensure VS Code uses the project virtual environment.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。