# arXiv MCP for AI Agents AI Agent Connect

> arXiv MCP. Access millions of scientific papers from arXiv. Search by author, category, or keyword and fetch metadata directly from the open-access archive.

## Overview
- **Category:** the-unthinkable
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_gHbFKVVkFdFXUhxEkwegYNZKOGLbcWlkvXYwdteW/ai-agent-connect
- **Tags:** scholarly-articles, academic-search, metadata-extraction, open-access, research-data, physics-math-cs

## Description

The arXiv MCP lets you pull the latest preprints and scholarly articles directly into your workspace. Imagine you're deep into a research project and need to see what's been published in machine learning over the last week. Instead of opening a new tab, navigating a complex archive, and manually copying titles into your notes, you can just ask your agent to do the heavy lifting. This Connector lets you grab specific papers by keyword or author, stay on top of new submissions in your niche using daily feeds, and even pull structured metadata for bigger projects like citation mapping. It's a huge time saver for anyone who spends hours digging through academic archives. By adding this arXiv MCP to your setup on Vinkius, you turn your AI into a research partner that can navigate the world's biggest open-access archive for you. It's about getting the information you need without the friction of manual searching. You can quickly check the latest work in physics, mathematics, or computer science without the usual friction of manual browsing. You can find a specific paper ID or identify the latest trends in a specific category. You get the ability to stay updated on the most recent preprints while keeping your focus on the actual analysis.

## Tools

### get_atom_feed
The get_atom_feed tool retrieves the daily Atom feed for a specific arXiv category or a combination of categories.

### oai_pmh_request
The oai_pmh_request tool requests raw XML metadata from the arXiv OAI-PMH interface for deep research analysis.

### get_rss_feed
The get_rss_feed tool pulls the daily RSS feed for specific arXiv categories to keep you updated on new submissions.

### search_articles
The search_articles tool searches for articles using keywords, authors, or a list of unique arXiv IDs.

## Prompt Examples

**Prompt:** 
```
Search arXiv for recent papers about 'Large Language Models' in the computer science category.
```

**Response:** 
```
I've found several recent papers on Large Language Models in the computer science category. Notable titles include:

* **Attention is All You Need**
* **Language Models are Few-Shot Learners**

Would you like me to fetch the full metadata for any of these?
```

**Prompt:** 
```
Get the daily RSS feed for the 'cs.AI' category to see today's new submissions.
```

**Response:** 
```
Fetching today's updates for Artificial Intelligence... I've retrieved the latest submissions, including:

* **Neural Network Optimization**
* **Reinforcement Learning Progress**

Should I summarize the top results for you?
```

**Prompt:** 
```
Use OAI-PMH to list the metadata formats available on arXiv.
```

**Response:** 
```
I've queried the OAI-PMH interface. arXiv supports several metadata formats including:

* **oai_dc**
* **arXiv**
* **arXivRaw**

Which format would you like to use for harvesting records?
```

## Capabilities

### Find papers by keyword
This action lets you find specific papers using keywords or authors.

### Get daily RSS feeds
This tool pulls daily updates for specific categories into your workspace.

### Get daily Atom feeds
This tool retrieves daily Atom feeds for various scientific domains.

### Get raw XML metadata
This tool requests raw XML from the arXiv OAI-PMH interface.

## Use Cases

### Finding specific research on LLMs
A student asks for recent papers on Large Language Models in computer science. The agent uses search_articles to find the top results and provides a summary.

### Monitoring new preprints daily
A researcher wants to see what's new in AI. They ask the agent to pull the daily RSS feed for the cs.AI category and list the titles.

### Harvesting metadata for citation maps
A data scientist needs to map citations for a set of IDs. They use oai_pmh_request to pull the structured metadata for all records at once.

### Quickly fetching paper details
Someone has a list of IDs and needs the metadata. They ask the agent to fetch the details for the list using search_articles.

## Benefits

- Stop manual searching by using search_articles to find specific papers by keyword or author instantly.
- Keep up with new research automatically using get_rss_feed to pull daily updates into your workspace.
- Access structured data for large-scale analysis using oai_pmh_request to get raw XML metadata.
- Track multiple interests at once by using get_atom_feed to combine various categories into one feed.
- Speed up literature reviews by using search_articles to fetch details for a list of specific IDs in one go.

## How It Works

The bottom line is you get direct access to the arXiv archive without the manual overhead of browsing or copy-pasting.

1. Subscribe to the arXiv MCP on Vinkius
2. Connect your AI client to the archive
3. Ask your agent to search for papers or pull specific feeds

## Frequently Asked Questions

**Can I use the arXiv MCP to find specific research papers?**
Yes, you can search for papers by keyword, author, or specific categories to find exactly what you need for your research.

**Does the arXiv MCP support different scientific fields?**
It covers the entire archive, including physics, mathematics, computer science, and many other scholarly disciplines.

**How does the arXiv MCP help with literature reviews?**
It allows you to quickly pull metadata and summaries for multiple papers at once, making it much faster to organize your sources.

**Can I get daily updates from the arXiv MCP?**
Yes, you can pull daily RSS or Atom feeds for specific categories to stay current on the latest preprints and submissions.

**Is the arXiv MCP good for data science projects?**
It's excellent for harvesting structured metadata to map trends, analyze citations, or gather data for machine learning research.

**Does the arXiv MCP work with my AI client?**
It works with any MCP-compatible client like Claude, Cursor, or Windsurf, giving your agent direct access to the archive.

**How can I search for papers by a specific author?**
Use the `search_articles` tool and include the author prefix in the `search_query` parameter, for example: `au:Einstein`.

**Can I get the latest daily updates for a specific category like AI?**
Yes! Use the `get_rss_feed` or `get_atom_feed` tools with the category code (e.g., `cs.AI` for Artificial Intelligence).

**Is it possible to fetch metadata for a specific list of arXiv IDs?**
Absolutely. Use the `search_articles` tool and provide the IDs as a comma-separated string in the `id_list` parameter.