Skip to content

PaSa

An academic paper search agent that reads and filters papers

Visit official site

What PaSa is

The agent is described as an academic paper search agent built on large language model technology, able to call search tools, read papers and screen references by itself. The material states it was optimised through reinforcement learning and evaluated on the RealScholarQuery dataset.

What you can do with it

  • Generate search queries and call search tools to collect papers
  • Read paper content and screen the most relevant literature
  • Extend the search through citation networks
  • Run several rounds of search and filtering to refine results

Application scenarios

  • A researcher looks for recent papers on machine learning
  • A student finds references for a graduation thesis
  • A research team screens key literature in a specific field

Getting started

  • Visit the official site or use the API interface
  • Enter a detailed academic search need such as a topic or keywords
  • Let the agent call search tools and produce a paper list
  • Review the ranked and filtered recommendations

Pros & cons

✓ What we like

  • Multi-round search and filtering refine the results
  • Citation networks widen the pool of related papers

! What to watch out for

  • The documented entry point is the official site or an API interface
  • Results are filtered and ranked by the system before review

FAQ

How was it trained?

The material states reinforcement learning with the synthetic AutoScholarQuery dataset.

What datasets are named?

The material names AutoScholarQuery and the real-world RealScholarQuery dataset.

Last reviewed: 2026-11-15

More AI search engine tools

View all →

How we review