research-ai

Welcome to the documentation for the research-ai project.

This site contains an overview, how to get started, usage examples, and a full API reference generated from the source code in agent/, bot/, and shared/.

Contents

• Getting Started
  • Installation
  • Configure environment
  • Initialize the database
  • Building the documentation
  • Running the project
  • Troubleshooting
  • Helpful links
• Usage
  • CLI entrypoints
  • Code structure
  • Telegram bot commands
  • Configuration reference
  • Typical workflow
  • Development
• Use Cases
  • Research assistants
  • Teams and communities
  • Product and engineering
  • Education
  • REST API automations
• Contributing
  • Guidelines
  • Project conventions
  • Dev setup
  • Quality checks
  • Documentation
• Modules
  • Agent Package
  • Bot Package
  • API Package
  • Shared Package

Project overview

research-ai provides an end-to-end research assistant that:

• Generates multiple search queries for a user task

• Retrieves candidates from multiple sources (arXiv, Google Scholar, PubMed, GitHub) and deduplicates them

• Ranks with BM25 over title + abstract

• Analyzes the top candidates (LLM-backed or heuristic)

• Decides whether to notify and formats a concise report for Telegram

You can interact via Telegram, or call the REST API to run the pipeline programmatically. See the pages above to learn how to install, run, and extend the project.