Introducing ArXiv AI – Natural Language Access to arXiv, Powered by Dappier

[Mohammed Amaan]

Hi everyone,

I’m excited to share a new tool we’ve built using the arXiv API called ArXiv AI — a natural language interface to access and explore over 2.4 million research papers across physics, CS, engineering, and beyond.

What It Does

• Accepts freeform natural language questions like:

  • “Summarize recent advancements in graph neural networks”

  • “List papers on quantum error correction after 2022”

• Uses OpenAI function calling to convert questions into structured API calls

• Retrieves relevant arXiv papers and formats results with summaries, descriptions, and source links

• Supports complex research and exploration use cases — great for fast literature reviews

How It Works

1. User submits a query in plain English

2. OpenAI interprets it and generates a function call to the arXiv API

3. The result is presented in a clear, helpful format with clickable arXiv links

Pricing

• $0.006 per query

• Pay-as-you-go model hosted on Dappier Marketplace

Tech Stack

• Python

• OpenAI

• arXiv API

• Deployed on Dappier

---

I'm sharing this here to offer it as a helpful tool for the community. I’ve just joined the mailing list and would love to get feedback, answer questions, or hear from others building similar research-oriented tools.

Thanks for the continued work on the arXiv API — it's an incredible foundation for tools like this!

Best regards,
Amaan

[Tom Morris]

On Tue, May 13, 2025 at 1:43 AM Mohammed Amaan <mohamme...@gmail.com> wrote:

I’m excited to share a new tool we’ve built using the arXiv API called ArXiv AI — a natural language interface to access and explore over 2.4 million research papers across physics, CS, engineering, and beyond.

Since this is using the ArXiv name, is it safe to assume that it's sponsored by or affiliated with ArXiv and/or Cornell University?

Tom

p.s. Those links don't go anywhere useful for providing more information. They immediately hit a login / account creation page.

[Jake Weiskoff]

This project is not affiliated with arXiv.org, nor Cornell University.

 

Note that there are legal uses of arXiv’s brand and trademark, as explicitly outlined in the API TOU:

 

https://info.arxiv.org/help/api/tou.html

 

-- 

Jake Weiskoff

Project Manager, arXiv.org

Cornell Tech

ja...@arxiv.org

 

 

--
You received this message because you are subscribed to the Google Groups "arXiv API Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to api+uns...@arxiv.org.
To view this discussion visit https://groups.google.com/a/arxiv.org/d/msgid/api/CAE9vqEFrG-4gVMr22QqU%3DJiXjhmM5SSAsdcGcYv-JL40UZa6Cw%40mail.gmail.com.

[Mohammed Amaan]

Hi Jake and team,

Thanks for your message and for maintaining such an essential resource for the research community.

We’ve now renamed the tool from ArXiv AI to Arxiv Search.

We’ve also updated our messaging to express our appreciation for arXiv’s open access and API — it’s an incredible foundation that enables tools like this, and we’re thankful for the interoperability it provides.

Feel free to head over to https://marketplace.dappier.com/marketplace/arxiv-search to see the agent in action.

Please let us know if there’s anything else we should do to remain in alignment with the API Terms of Use.

Best regards,
Amaan
(on behalf of the Dappier team)

[Mohammed Amaan]

Hi Tom,

Thanks for the question — to clarify, this tool is not affiliated with or sponsored by arXiv.org or Cornell University. It’s an independent project built using the publicly available arXiv API, in line with the API Terms of Use.

You can try the tool here:
👉 https://marketplace.dappier.com/marketplace/arxiv-search

Access to the agent requires a Dappier account — this is intentional and necessary to support authentication, user-level rate limiting, and usage-based billing. We currently don’t have plans to offer unauthenticated or public access.

Appreciate your interest, and happy to answer any other questions.

Best regards,
Amaan
(on behalf of the Dappier team)

[Jake Weiskoff]

There is significant concern that you’re going to cause confusion with a name such as this, as users will not understand that your tool is not arXiv’s search. Please review:

 

https://info.arxiv.org/brand/index.html

 

for arXiv’s brand guidelines, specifically: https://info.arxiv.org/brand/brand-guidelines.html

 

-- 

Jake Weiskoff

Project Manager, arXiv.org

Cornell Tech

ja...@arxiv.org

 

 

From: Mohammed Amaan <mohamme...@gmail.com>
Date: Thursday, May 15, 2025 at 5:22 AM
To: arXiv API Discussion <a...@arxiv.org>
Cc: tfmo...@gmail.com <tfmo...@gmail.com>, kr...@dappier.com <kr...@dappier.com>, si...@ai20labs.com <si...@ai20labs.com>
Subject: Re: [arXiv-api] Introducing ArXiv AI – Natural Language Access to arXiv, Powered by Dappier

--

You received this message because you are subscribed to the Google Groups "arXiv API Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to api+uns...@arxiv.org.

To view this discussion visit https://groups.google.com/a/arxiv.org/d/msgid/api/1d98a917-5f55-4641-bebb-6db05f40e4can%40arxiv.org.

[Tom Morris]

Jake - it might be worth suggesting to your legal beagles the addition of some product naming examples to your branding guide.

Amaan - Trademark lawyers generally dislike (ie prohibit) use of trademarks as part of a compound product name for an unrelated organization, so "Amazing Search powered by ArXiv and Dappier" is generally acceptable, but "ArXiv Search" or "Dappier Search" isn't unless produced by the organization which owns the trademark.

Tom

[Krish Arvapally]

Hey Jake and Tom,

Appreciate the guidance on this and feedback! Totally understand and we want to ensure we are complying with the guidelines. 

I was about to the same question - what would be a good example name for this model. 

Would this be OK to say: 

Research Papers Search (powered by arXiv) 

Provides instant access to over 2.4 million open-access scholarly articles across domains including physics, mathematics, computer science, quantitative biology, quantitative finance, statistics, electrical engineering and systems science, and economics. This AI-powered search model is powered by arXiv and hosted by Dappier. 

BTW we are huge supporter of the work you guys are doing. 

Thanks,

Krish

Sent via Superhuman

[Jake Weiskoff]

Dear Krish,

 

The name “arXiv” should not appear in the name of your product. You are welcome to acknowledge arXiv within the subtext of your product, but you cannot use “arXiv” as part of your name.  

 

-- 

Jake Weiskoff

Project Manager, arXiv.org

Cornell Tech

ja...@arxiv.org

 

 

From: Krish Arvapally <kr...@dappier.com>
Date: Thursday, May 15, 2025 at 11:46 AM
To: Tom Morris <tfmo...@gmail.com>
Cc: Jake Weiskoff <ja...@arxiv.org>, Mohammed Amaan <mohamme...@gmail.com>, a...@arxiv.org <a...@arxiv.org>, si...@ai20labs.com <si...@ai20labs.com>
Subject: Re: [arXiv-api] Introducing ArXiv AI – Natural Language Access to arXiv, Powered by Dappier

[Krish Arvapally]

Hi Jake,

How's the following: 

Research Papers Search

Provides instant access to over 2.4 million open-access scholarly articles across domains including physics, mathematics, computer science, quantitative biology, quantitative finance, statistics, electrical engineering and systems science, and economics. This AI-powered search model is powered by arXiv and hosted by Dappier.  Thank you to arXiv for use of its open access interoperability. 

Thanks,

Krish 

Sent via Superhuman

On Tue, May 20, 2025 at 10:55 AM, Jake Weiskoff <ja...@arxiv.org> wrote:

Dear Krish,

 

The name “arXiv” should not appear in the name of your product. You are welcome to acknowledge arXiv within the subtext of your product, but you cannot use “arXiv” as part of your name.  

 

-- 

Jake Weiskoff

Project Manager, arXiv.org

Cornell Tech

jake@arxiv.org

 

 

From: Krish Arvapally <krish@dappier.com>
Date: Thursday, May 15, 2025 at 11:46 AM
To: Tom Morris <tfmorris@gmail.com>
Cc: Jake Weiskoff <jake@arxiv.org>, Mohammed Amaan <mohammedamaan79@gmail.com>, api@arxiv.org <api@arxiv.org>, siva@ai20labs.com <siva@ai20labs.com>
Subject: Re: [arXiv-api] Introducing ArXiv AI – Natural Language Access to arXiv, Powered by Dappier

Hey Jake and Tom,

 

Appreciate the guidance on this and feedback! Totally understand and we want to ensure we are complying with the guidelines. 

I was about to the same question - what would be a good example name for this model. 

 

Would this be OK to say: 

 

Research Papers Search (powered by arXiv) 

Provides instant access to over 2.4 million open-access scholarly articles across domains including physics, mathematics, computer science, quantitative biology, quantitative finance, statistics, electrical engineering and systems science, and economics. This AI-powered search model is powered by arXiv and hosted by Dappier. 

 

BTW we are huge supporter of the work you guys are doing. 

 

Thanks,

Krish

 

Sent via Superhuman

 

[Jake Weiskoff]

The “Powered by arXiv” may imply a relationship that doesn’t exist, so we would recommend that you not use it in your blurb. I would recommend something more like

 

… This AI-powered search model is based upon arXiv and hosted by Dappier.  Thank you to arXiv for use of its open access interoperability. 

 

These statements are factual, which is better than “Powered by”, which could imply either that you’re physically running it on our servers (and therefore arXiv would be responsible for its maintenance), or that there is some form of financial relationship between our entities (neither of which are correct, to my knowledge).

 

-- 

Jake Weiskoff

Project Manager, arXiv.org

Cornell Tech

ja...@arxiv.org

 

 

From: Krish Arvapally <kr...@dappier.com>
Date: Tuesday, May 20, 2025 at 11:40 AM
To: Jake Weiskoff <ja...@arxiv.org>
Cc: Tom Morris <tfmo...@gmail.com>, Mohammed Amaan <mohamme...@gmail.com>, a...@arxiv.org <a...@arxiv.org>, si...@ai20labs.com <si...@ai20labs.com>
Subject: Re: [arXiv-api] Introducing ArXiv AI – Natural Language Access to arXiv, Powered by Dappier

Hi Jake,

 

How's the following: 

 

Research Papers Search

Provides instant access to over 2.4 million open-access scholarly articles across domains including physics, mathematics, computer science, quantitative biology, quantitative finance, statistics, electrical engineering and systems science, and economics. This AI-powered search model is powered by arXiv and hosted by Dappier.  Thank you to arXiv for use of its open access interoperability. 

 

Thanks,

Krish 

 

Sent via Superhuman

 

 

On Tue, May 20, 2025 at 10:55 AM, Jake Weiskoff <ja...@arxiv.org> wrote:

Dear Krish,

 

The name “arXiv” should not appear in the name of your product. You are welcome to acknowledge arXiv within the subtext of your product, but you cannot use “arXiv” as part of your name.  

 

-- 

Jake Weiskoff

Project Manager, arXiv.org

Cornell Tech

ja...@arxiv.org

 

 

From: Krish Arvapally <kr...@dappier.com>
Date: Thursday, May 15, 2025 at 11:46 AM

To: Tom Morris <tfmo...@gmail.com>
Cc: Jake Weiskoff <ja...@arxiv.org>, Mohammed Amaan <mohamme...@gmail.com>, a...@arxiv.org <a...@arxiv.org>, si...@ai20labs.com <si...@ai20labs.com>

Subject: Re: [arXiv-api] Introducing ArXiv AI – Natural Language Access to arXiv, Powered by Dappier

[Krish Arvapally]

Ah perfect, thank you for the guidance :) We were scratching our heads to figure out the right way. 

We'll update to: 

Research Papers Search

Provides instant access to over 2.4 million open-access scholarly articles across domains including physics, mathematics, computer science, quantitative biology, quantitative finance, statistics, electrical engineering and systems science, and economics. This AI-powered search model is based upon arXiv and hosted by Dappier.  Thank you to arXiv for use of its open access interoperability. 

Thanks,

Krish

Sent via Superhuman

On Tue, May 20, 2025 at 11:50 AM, Jake Weiskoff <ja...@arxiv.org> wrote:

The “Powered by arXiv” may imply a relationship that doesn’t exist, so we would recommend that you not use it in your blurb. I would recommend something more like

 

… This AI-powered search model is based upon arXiv and hosted by Dappier.  Thank you to arXiv for use of its open access interoperability. 

 

These statements are factual, which is better than “Powered by”, which could imply either that you’re physically running it on our servers (and therefore arXiv would be responsible for its maintenance), or that there is some form of financial relationship between our entities (neither of which are correct, to my knowledge).

 

-- 

Jake Weiskoff

Project Manager, arXiv.org

Cornell Tech

jake@arxiv.org

 

 

From: Krish Arvapally <krish@dappier.com>
Date: Tuesday, May 20, 2025 at 11:40 AM
To: Jake Weiskoff <jake@arxiv.org>
Cc: Tom Morris <tfmorris@gmail.com>, Mohammed Amaan <mohammedamaan79@gmail.com>, api@arxiv.org <api@arxiv.org>, siva@ai20labs.com <siva@ai20labs.com>
Subject: Re: [arXiv-api] Introducing ArXiv AI – Natural Language Access to arXiv, Powered by Dappier

Hi Jake,

 

How's the following: 

 

Research Papers Search

Provides instant access to over 2.4 million open-access scholarly articles across domains including physics, mathematics, computer science, quantitative biology, quantitative finance, statistics, electrical engineering and systems science, and economics. This AI-powered search model is powered by arXiv and hosted by Dappier.  Thank you to arXiv for use of its open access interoperability. 

 

Thanks,

Krish 

 

Sent via Superhuman

 

 

On Tue, May 20, 2025 at 10:55 AM, Jake Weiskoff <jake@arxiv.org> wrote:

Dear Krish,

 

The name “arXiv” should not appear in the name of your product. You are welcome to acknowledge arXiv within the subtext of your product, but you cannot use “arXiv” as part of your name.  

 

-- 

Jake Weiskoff

Project Manager, arXiv.org

Cornell Tech

jake@arxiv.org

 

 

From: Krish Arvapally <krish@dappier.com>
Date: Thursday, May 15, 2025 at 11:46 AM

To: Tom Morris <tfmorris@gmail.com>
Cc: Jake Weiskoff <jake@arxiv.org>, Mohammed Amaan <mohammedamaan79@gmail.com>, api@arxiv.org <api@arxiv.org>, siva@ai20labs.com <siva@ai20labs.com>

Subject: Re: [arXiv-api] Introducing ArXiv AI – Natural Language Access to arXiv, Powered by Dappier