Skip to content
/ Reportr Public

An Agentic AI to search for Research Papers

Notifications You must be signed in to change notification settings

ms1963/Reportr

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

12 Commits
 
 
 
 
 
 

Repository files navigation

Reportr - An Agentic/Multi-Agent AI System

image

Motivation

The motivation of Reportr was to design and implement an Agentic AI system. The system is extensible. What I implemented is a tool that runs either in interactive mode (CLI-based) or in query mode. The user specifies one or multiple research topics. The agent searches different sources like ArVix or Google Scholar for current, high-rated research papers, articles, blog-posts. It concurrrently downloads up to a maximum number of files. Then it rates and summarizes them. The research results are eventually shown on the screen and also stored in an .md document.

It is possible to configure Reportr to run periodically using time schedules like "all two week on Monday at 8:00". The documents are partitioned in chunks and stored in a vector database. So far Reportr is not using this database, but can be easily extended with a query component that allows to ask questions about the stored content.

You may use Reportr with local or cloud-based LLMs. For local inference llama.cpp and Ollama are supported, but it is easy to add other inference engines as well.

About

An Agentic AI to search for Research Papers

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages