Do I need an arXiv account?

No. This connector needs no authentication at all, which makes it the quickest way to see what a research-capable agent does with a real corpus.

Can it read the full paper or just the abstract?

It works from the paper itself, not only the abstract — which is what makes “explain the method, then say what it does not establish” a question worth asking.

Will it make up citations?

It works from papers it actually retrieved, and cites them. That is a much stronger position than a model answering from memory, though a claim that matters is still worth opening the source to check.

Research & academic

arXiv + Burrak AI

Search and read the arXiv preprint corpus — no account, no sign-in.

No sign-in required

arXiv is where most machine-learning and physics work appears before it appears anywhere else. This connector needs no authentication at all, so an agent can search, fetch and read papers immediately — which makes it the cheapest possible way to see what a research-capable agent does with a real corpus.

In the connector catalogue: Search preprints, read abstracts, fetch paper metadata.

Things you could ask for

Written the way you would actually say them. An agent plans the steps itself.

“Find the most cited papers on retrieval-augmented generation from the past six months and summarise what changed.”
“Read this paper and explain its method in plain language, then say what it does not establish.”
“Every Monday, check for new preprints on agent evaluation and send me anything genuinely novel.”
“Compare the approaches in these three papers and lay out where they disagree.”
“Trace the lineage of this idea back through the papers it builds on.”

No sign-in required — this one works the moment you enable it.

Connecting it

1Create a Burrak account — the free plan is enough to try this.
2Open Connectors and choose arXiv.
3That is the whole setup — this connector needs no account of its own.
4Give an agent a task. It will use the connector when the work calls for it.

Start free How connectors work

arXiv and Burrak, in practice

Do I need an arXiv account?: No. This connector needs no authentication at all, which makes it the quickest way to see what a research-capable agent does with a real corpus.
Can it read the full paper or just the abstract?: It works from the paper itself, not only the abstract — which is what makes “explain the method, then say what it does not establish” a question worth asking.
Will it make up citations?: It works from papers it actually retrieved, and cites them. That is a much stronger position than a model answering from memory, though a claim that matters is still worth opening the source to check.