# arXiv MCP for AI Agents AI Agent Connect

> arXiv MCP lets your AI agent search over 2.5 million scientific preprints across physics, computer science, and math. It pulls full abstracts, author lists, and direct PDF links so you can stay on top of the latest research without manually browsing the site.

## Overview
- **Category:** the-unthinkable
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_GMLyHgp9f43T21mrifi7wgMeTSOXEgTkFrJKs2xB/ai-agent-connect
- **Tags:** preprint-search, scientific-research, metadata-retrieval, academic-database, document-retrieval

## Description

You're trying to keep up with the latest breakthroughs in a field that moves faster than peer review can keep up with. Instead of manually digging through categories or bookmarking dozens of tabs, this Connector lets your AI agent do the heavy lifting. You can ask your agent to find the most recent papers on specific topics like diffusion models or quantum error correction and get a curated list of results with abstracts and direct links to the PDFs. It handles the filtering across different scientific domains, whether you're looking for a specific math theorem or a new biology preprint. This is a core part of the Vinkius catalog designed to bridge the gap between raw academic data and your daily research workflow. You get the ability to pull deep metadata on specific papers by their ID, ensuring you have the full context, like authors, publication dates, and DOIs, right in your chat interface.

## Tools

### search_arxiv
Search 2.5 million preprints by keyword or category to find relevant research papers. It returns titles, authors, and direct PDF links for your results.

### get_arxiv_paper
Retrieve full metadata for a specific paper using its unique arXiv ID. This gives you the abstract, publication date, and DOI in one go.

## Prompt Examples

**Prompt:** 
```
Find the latest papers on large language models and reasoning.
```

**Response:** 
```
### Recent LLM Reasoning Preprints

I found several recent papers focusing on reasoning capabilities. Here are the top results:

| Title | Authors | Category | Link |
| :--- | :--- | :--- | :--- |
| **Chain-of-Thought Prompting Elicits Reasoning** | Wei et al. | cs.CL | [PDF](https://arxiv.org/pdf/2201.11903) |
| **Tree-of-Thought Prompting** | et al. | cs.AI | [PDF](https://arxiv.org/pdf/2305.10601) |

**Quick Summary:**
- Most papers focus on multi-step reasoning.
- Several explore self-consistency checks.
- Many provide new benchmarks for mathematical logic.
```

**Prompt:** 
```
Get the full details of the 'Attention Is All You Need' paper from arXiv.
```

**Response:** 
```
### Paper Details: arXiv:1706.03762

**Title:** Attention Is All You Need
**Authors:** Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Łukasz Kaiser, Illia Polosukhin
**Published:** 2017-06-12
**Categories:** cs.CL, cs.LG

**Abstract:**
The dominant sequence transduction models are based on complex recurrent or convolutional neural networks. We propose a new simple network architecture, the Transformer, based solely on attention mechanisms. 

📥 **Direct PDF:** [arxiv.org/pdf/1706.03762](https://arxiv.org/pdf/1706.03762)
```

**Prompt:** 
```
Search arXiv for papers on quantum error correction in the physics category.
```

**Response:** 
```
### Quantum Error Correction Research (Physics)

I found several relevant preprints in the **quant-ph** category:

*   **Suppressing quantum errors by scaling a surface code logical qubit** - Demonstrates below-threshold error rates using Google's Sycamore processor.
*   **Topological codes and Floquet codes** - Recent explorations in error correction.
*   **Bosonic qubit approaches** - New methodologies in hardware-efficient correction.

Would you like me to pull the full abstract for any of these?
```

## Capabilities

### Search millions of preprints
Find research papers by keyword or category across multiple scientific domains.

### Pull full abstracts
Get the complete summary and author list for any specific paper ID.

### Get direct PDF links
Receive direct download links for every search result.

### Filter by scientific domain
Narrow down results to specific categories like physics, math, or cs.AI.

### Retrieve paper metadata
Get publication dates, DOIs, and categories for specific arXiv IDs.

## Use Cases

### Finding recent LLM research
A researcher asks for the latest work on reasoning capabilities. The agent uses search_arxiv to find 10 recent preprints and summarizes them.

### Retrieving specific paper details
A student wants the abstract for a specific ID. The agent uses get_arxiv_paper to show the full abstract and author list immediately.

### Filtering physics preprints
An engineer wants papers on quantum error correction. The agent filters the physics category using search_arxiv and lists results.

### Summarizing math theorems
A mathematician looks for new theorems from last week. The agent searches the math category and provides a summary of new preprints.

## Benefits

- Stop manually browsing different categories on the arXiv website to find what you need.
- Use search_arxiv to get a curated list of papers based on specific keywords.
- Get direct download links for every result to save time on manual downloads.
- Pull complete abstracts with get_arxiv_paper to decide what to read first.
- Access research in fields like cs.AI and physics without leaving your chat interface.
- Keep your research organized by pulling metadata like DOIs and publication dates.

## How It Works

The bottom line is you get instant access to the world's largest preprint database through your AI chat.

1. Subscribe to the arXiv MCP on Vinkius.
2. Connect your preferred AI client to the catalog.
3. Ask your agent to search for specific research topics or paper IDs.

## Frequently Asked Questions

**Can the arXiv MCP find papers in my specific field?**
Yes, it covers physics, computer science, math, biology, and economics. You can search by keyword or use specific category filters like cs.AI or physics.

**Does the arXiv MCP provide direct links to the papers?**
Yes, every search result includes a direct link to the PDF so you can read the full research immediately.

**Do I need an API key to use the arXiv MCP?**
No, you don't need any special keys. Just connect the Connector through Vinkius and start searching.

**Can I get the abstract for a specific paper ID?**
Yes, you can provide any arXiv ID and the agent will pull the full abstract, author list, and publication date for you.

**Is this Connector good for finding the very latest research?**
It's perfect for that. It accesses preprints, which are often released before they are officially published in journals.

**How does the arXiv MCP handle different scientific categories?**
It supports specific categories like cs.AI, physics, and math, allowing you to narrow down your search results significantly.

**What is a preprint and how does arXiv work?**
A preprint is a scientific paper shared publicly before formal peer review. arXiv allows researchers to share their findings immediately upon completion, accelerating scientific communication. Papers are assigned a permanent arXiv ID (e.g., 2106.09685) and are freely accessible forever. Many landmark papers in AI, physics, and math appeared on arXiv months or years before journal publication.

**What scientific categories and disciplines does arXiv cover?**
arXiv covers 8 major domains: Physics (astro-ph, cond-mat, hep, quant-ph), Mathematics (math.*), Computer Science (cs.AI, cs.LG, cs.CL, cs.CV, cs.CR, etc.), Quantitative Biology (q-bio), Quantitative Finance (q-fin), Statistics (stat), Electrical Engineering (eess), and Economics (econ). Each domain has multiple sub-categories for precise filtering.

**Is arXiv free and do I need an API key?**
Yes, arXiv is completely free and operated as a non-profit by Cornell University. No API key or registration is required for search queries. The only limitation is a rate limit of approximately 1 request per 3 seconds for the search API. All papers are freely downloadable as PDF and accessible in perpetuity.