You will run and use an MCP server designed to monitor AI and LLM research in real time. This server collects data from multiple sources, lets you search and summarize progress, and caches results to stay responsive while minimizing API usage. It is intended to be connected through an MCP client so you can ask questions and receive structured, up-to-date insights about AI research topics.

How to use

Connect to the MCP server through your MCP client. You have two local startup options that the server supports: run the server with a Python command or run it using a specific Python executable path. In either case, the MCP server exposes a named connection you can select in your client (ai-research). Once connected, you can ask for the latest papers, trending GitHub repositories, daily or weekly summaries, or domain-specific searches.

How to install

Prerequisites: Python 3.10 or newer, and the pip package manager. You also need an MCP client such as Claude Desktop to connect to the server.

Step-by-step commands to install and run the MCP server locally:

# 1. Clone the project repository
# (Use the repository you have access to)
git clone https://github.com/nanyang12138/AI-Research-MCP.git
cd AI-Research-MCP

# 2. Install the Python package in editable mode
pip install -e .

# 3. (Optional) Create and configure a GitHub token in an environment file
cp .env.example .env
# Edit .env to add your GitHub Token

If you prefer, you can start the server using a Python interpreter directly (see the two provided startup options under Claude Desktop configuration)."}]} ,{

## Configuration and startup options

Two explicit MCP startup configurations are provided to run the server locally. Use the one that matches your environment. Each configuration starts the server via Python and sets the required environment variable for GitHub API access.

{ "mcpServers": { "ai-research": { "command": "python", "args": ["-m", "ai_research_mcp.server"], "env": { "GITHUB_TOKEN": "your_github_token_here" } } } }

## Additional configuration details

Environment variables control caching behavior and API access. Create a .env file to customize these values. The server uses a local cache directory by default and respects expiry times for different data sources.

GitHub Personal Access Token (strongly recommended)

Increases API rate limits to 5000 requests/hour when provided

GITHUB_TOKEN=ghp_your_token_here

Cache directory (optional, defaults to .cache)

CACHE_DIR=.cache

Cache expiry times (in seconds)

CACHE_EXPIRY_GITHUB=3600 # 1 hour CACHE_EXPIRY_ARXIV=7200 # 2 hours CACHE_EXPIRY_SUMMARY=86400 # 24 hours

## Troubleshooting

If you encounter issues, check common causes such as missing or invalid tokens, incorrect file paths, or environment variable configuration problems. Ensure the MCP client shows the connected status (a keyboard-like icon or similar indicator) after starting the server.

Example checks you can perform locally

python -m ai_research_mcp.server --help

## Usage patterns you can perform with the MCP server

You can perform domain-specific searches, fetch the latest papers, find trending GitHub repositories, and generate daily or weekly summaries. These actions are exposed as MCP tools and can be invoked through your MCP client interface by selecting the ai-research server and issuing natural language prompts.

## Available tools

### search\_latest\_papers

Search the most recent AI papers from multiple sources such as arXiv and Papers with Code.

### search\_github\_repos

Search trending or high-star AI GitHub repositories.

### get\_daily\_papers

Retrieve today’s featured papers and summaries pulled from multiple sources.

### get\_trending\_repos

Fetch GitHub repositories that are currently trending in AI topics.

### get\_trending\_models

Fetch Hugging Face models that are trending or popular.

### search\_by\_area

Search AI research by domain or area (LLM, Vision, Robotics, etc.).

### generate\_daily\_summary

Generate a daily digest of AI research progress.

### generate\_weekly\_summary

Generate a weekly digest of AI research progress.