scholaraio  by ZimoLiao

AI-native research terminal for AI agents

Created 3 weeks ago

New!

258 stars

Top 98.0% on SourcePulse

GitHubView on GitHub
Project Summary

Summary

ScholarAIO provides a knowledge infrastructure for AI agents, transforming research papers into a structured, searchable format. It enables agents to perform advanced tasks like reproducing methods, verifying claims, and drafting manuscripts, streamlining the AI-powered research loop.

How It Works

ScholarAIO parses PDFs into structured Markdown, preserving figures and LaTeX, then enriches metadata via APIs. It employs hybrid search (keyword FTS5 + semantic FAISS) for retrieval and integrates with AI agents via standardized skills or direct repo usage, allowing agents to leverage the processed knowledge base.

Quick Start & Requirements

Install via git clone, pip install -e ".[full]", and scholaraio setup, then run claude in the repo root. Alternatively, use as a Claude Code plugin. Requires Python 3.10+, an LLM API key (e.g., Anthropic, OpenAI, Google, DeepSeek), and optionally a MINERU_API_KEY. An embedding model (~1.2 GB) auto-downloads. See docs/getting-started/agent-setup.md for details.

Highlighted Details

  • PDF Processing: Converts PDFs to Markdown, retaining figures, LaTeX, and supporting long documents.
  • Hybrid Search: Fuses keyword (FTS5) and semantic (FAISS) indexing with RRF ranking.
  • Topic Discovery: BERTopic for automatic clustering and interactive visualization.
  • Multi-Source Import: Supports Endnote (XML/RIS) and Zotero (API/SQLite).
  • AI-Assisted Writing: Features for literature review, drafting, citation checking, and gap analysis.
  • Persistent Notes: Saves agent analysis per paper to avoid redundant processing.

Maintenance & Community

Licensed under MIT, the project appears primarily maintained by Zi-Mo Liao. No specific community channels or external sponsorships are detailed.

Licensing & Compatibility

The MIT License is permissive for commercial use and integration into closed-source projects.

Limitations & Caveats

Full functionality requires LLM API keys. Advanced PDF parsing relies on the optional MINERU_API_KEY or self-hosting MinerU. Agent integration methods vary by agent type.

Health Check
Last Commit

3 days ago

Responsiveness

Inactive

Pull Requests (30d)
35
Issues (30d)
10
Star History
258 stars in the last 25 days

Explore Similar Projects

Feedback? Help us improve.