This project provides an AI-driven "skill" for Claude Code, designed to streamline academic paper research across multiple disciplines and platforms. It targets researchers and power users seeking efficient discovery of cutting-edge papers, citation tracking, and reliable open-access PDF acquisition, ultimately redefining the academic search experience.

How It Works

Academic-Search employs a multi-platform, multi-disciplinary approach, integrating sources like arXiv, Semantic Scholar, OpenAlex, Crossref, Unpaywall, Google Scholar, and CNKI. Its core strategy involves a two-pass search: initially presenting a lightweight summary table before fetching full metadata upon user confirmation, or directly outputting results if a specific quantity is requested. Query expansion automatically generates complementary queries to boost coverage by 30-50%. Ranking prioritizes recency (last 6 months) over citation count to surface the latest advancements, with CCF venue rankings used as a reference. It merges results from various platforms using DOI/arXiv ID and handles open-access PDF acquisition by constructing direct links and annotating access statuses (open_pdf, needs_institution, etc.) without bypassing paywalls. For platforms lacking APIs or employing strong anti-scraping measures like Google Scholar and CNKI, it utilizes a Chrome DevTools Protocol (CDP) browser mode, leveraging the user's logged-in session.

Quick Start & Requirements

• Installation: Manual installation involves git clone https://github.com/Mingyue-Cheng/academic-search ~/.claude/skills/academic-search. Alternatively, it can be installed via Claude or linked locally for development.
• Prerequisites: The CDP browser mode requires Chrome remote debugging to be enabled. API platforms must be directly accessible.
• Links: GitHub Repository

Highlighted Details

• Multi-Platform & Multi-Disciplinary: Integrates numerous academic databases and tailors search strategies for CS/AI, Medicine, Physics, Chemistry, Social Sciences, and Humanities.
• Forward-Looking Ranking: Prioritizes recency (last 6 months) to highlight cutting-edge research.
• Reliable OA PDF Handling: Generates manifests and downloads only legally open PDFs, clearly annotating access status and restrictions.
• CDP Browser Mode: Enables access to anti-bot platforms like Google Scholar and CNKI by integrating with a user's Chrome instance.

Maintenance & Community

The project is authored by Mingyue Cheng. No specific community channels (Discord, Slack) or details on broader contributor activity, sponsorships, or partnerships are provided in the README.

Licensing & Compatibility

The project is released under the MIT License, which is permissive and generally suitable for commercial use and integration into closed-source projects.

Limitations & Caveats

This tool strictly adheres to accessing legally open-access PDFs and does not bypass paywalls, Sci-Hub, WebVPN, or Tor. Access to certain platforms like Google Scholar and CNKI is dependent on the setup of the CDP browser mode. The project differentiates itself by focusing on search and metadata enrichment rather than aggressive, unrestricted PDF downloading. Users are advised to obtain a free S2 API key to mitigate rate limiting issues.