Tagny MCP Server
Enables web browsing capabilities for locally served LLMs through URL text fetching, link extraction, and web search using Brave and DuckDuckGo engines. Designed to enhance LLMs with real-time web access through the MCP protocol.
README
tagny-mcp-server
An MCP server with web search, URL text fetching, and more tools to enhance locally served LLMs.
Features
- Fetch plain text content from URLs
- Extract all links from a webpage
- Designed to work with locally served LLMs via MCP protocol
Installation
- from source
uv pip install -e .
# with dev dependencies
uv sync
- from PyPI repository
pip install tagny-mcp-server
Usage
Start the MCP server:
- from source code:
uv run -m tagny_mcp_server
- after proper installation:
- as described in the pyproject.toml at
project.scripts, you can run the with the default arguments :
- as described in the pyproject.toml at
tagny-mcp-server
- customize the arguments by running like this for example:
# see the help
uv run -m tagny_mcp_server --help
# see the version
uv run -m tagny_mcp_server --version
# change the default port
uv run -m tagny_mcp_server --port 5002
The server will run using Server-Sent Events (SSE) transport.
Tools
fetch_url_text
Downloads and parses HTML content from a URL, returning only the visible text.
fetch_page_links
Returns a list of all hyperlinks found on a webpage.
search_web_with_brave
An MCP tool that performs web searches using Brave search engine
search_web_with_duckduckgo
An MCP tool that performs web searches using DuckDuckGo
Project Structure
src/tagny_mcp_server/__init__.py- Main package initializer that exports the versionsrc/tagny_mcp_server/__main__.py- Entry point that launches the MCP server with command-line argumentssrc/tagny_mcp_server/__version__.py- Contains the package version stringsrc/tagny_mcp_server/config.py- Configures the FastMCP server instance with name, instructions, and versionsrc/tagny_mcp_server/web_access/__init__.py- Package initializer for web access tools that exports URL text fetching functionssrc/tagny_mcp_server/web_access/url_text_fetcher.py- Implements tools for fetching URL text and extracting links from web pagessrc/tagny_mcp_server/web_access/web_search.py- Implements Brave and DuckDuckGo search toolspyproject.toml- Project metadata, dependencies, and build configurationtests/scripts/client.py- Example client script for testing the servertests/test_web_access_tools.py- Tests for the web access tools.gitignore- Git ignore rulesREADME.md- Project documentationLICENSE- License informationCHANGELOG.md- Release notes and version historyCONTRIBUTING.md- Guidelines for contributing to the project.pre-commit-config.yaml- Pre-commit hooks configuration.python-version- Python version specification.bumpversion.toml- Configuration for version bumping tool.github/- GitHub workflow and issue template files
Dependencies
beautifulsoup4- HTML parsingfastmcp- MCP server frameworkrequests- HTTP requests
Testing
Tests are located in tests/ and can be run with pytest:
pytest
Example client usage is shown in scripts/client.py.
License
This project is licensed under the MIT License - see the LICENSE file for details.
Contributing
Contributions are welcome! Please read the CONTRIBUTING.md file for details on our code of conduct and the process for submitting pull requests.
Changelog
See CHANGELOG.md for a history of changes to this project.
Code Quality
This project uses pre-commit hooks to maintain code quality. Install them with:
pre-commit install
The following tools are used:
- black (code formatting)
- flake8 (linting)
- isort (import sorting)
- detect-secrets (secret detection)
- ruff (linting and fixing)
For more information, see .pre-commit-config.yaml.
CI/CD
This project uses GitHub Actions for continuous integration and deployment:
- Build workflow - runs tests and code quality checks
- Publish to Docker Hub - publishes the Docker image on release
- Publish to PyPI - publishes to PyPI on release
For more information, see .github/workflows.
Running with Docker
You can build and run the Docker container using the provided Dockerfile. To test locally, you can use act following the instructions at https://nektos.github.io/act/.
# Build the Docker image
docker build -t tagny-mcp-server .
# Run the Docker container
docker run -p 8000:8000 tagny-mcp-server
This will start the MCP server inside a Docker container, accessible at http://localhost:8000/sse.
That you can integrate in a mcp.json file like this:
{
"mcpServers": {
"tagny-mcp-server": {
"url": "http://localhost:8000/sse"
}
}
}
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。
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。
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。
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。
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。
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。
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。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。