中文 | English

Your AI Research Copilot

Stop searching papers. Start understanding research.

PaperLens is an AI research copilot that turns the messy, fragmented process of academic literature work into a single, coherent workflow. It's like Cursor, but for research papers.

Most researchers spend their days bouncing between Google Scholar, arXiv, PDF readers, and half a dozen browser tabs. PaperLens replaces all of that with one tool:

Find → Filter → Read → Understand → Save → Cite

• 43 academic databases searched in parallel, including PubMed, Scopus, ScienceDirect, arXiv, CNKI, and more
• 300+ Chinese universities accessible via CARSI — no VPN needed for ScienceDirect, Scopus, JSTOR from anywhere
• 100% local-first — all data stays on your machine. Runs fully offline with Ollama. No account required.

Search & Discover	Paper List	AI Deep Analysis

Multi-Paper Compare	Citation Graph

Google Scholar → arXiv → open PDF → not what I need → keep searching → open more tabs → 20 tabs later, still no clarity.

Describe your research question in natural language. PaperLens searches 43 databases at once, ranks results by relevance, and lets you:

• Read a paper in 2 minutes instead of 30 — AI extracts what matters: method, innovation, strengths, weaknesses, future directions
• Find the right paper in seconds, not hours — AI understands your research intent, not just keyword matching
• Compare papers side-by-side — spot contradictions, identify research gaps, find complementary approaches
• Discover connections — interactive citation graph reveals how papers relate, who cites whom, and emerging research clusters

PaperLens isn't just a search tool. It's a complete pipeline:

Describe your research interest in natural language. PaperLens AI constructs the optimal query and searches 43 academic databases in parallel — PubMed, OpenAlex, CrossRef, arXiv, Semantic Scholar, ScienceDirect, Scopus, JSTOR, CNKI, Wanfang, VIP, Google Scholar, and more. Results are deduplicated, enriched, and ranked in seconds.

Narrow down results by journal, year, field tags, publication type, or metadata completeness. Every paper gets a metadata score (0-100%) so you know what's well-documented and what's missing data.

Select any paper. AI extracts the essence:

Summary  →  Deep Analysis  →  Method  →  Innovation  →  Strength  →  Weakness  →  Future Work  →  Related Papers

This isn't a one-sentence summary. It's a 7-dimension structured breakdown — like having a senior colleague walk you through the paper.

Compare multiple papers side-by-side. AI identifies:

• Where approaches agree and diverge
• Research gaps no one has addressed
• Contradictions between findings
• Promising directions for future work

Export in any format: RIS, BibTeX, CSV, EndNote XML. Generate citations in APA 7th, MLA 9th, GB/T 7714, Chicago 17th, Vancouver. Track reading status (unread / reading / read). Sync tags bidirectionally with Zotero. Download OA PDFs with one click.

• 43 data sources searched in parallel — see the full list in ROADMAP.md
• Natural language — Chinese or English, AI understands your intent
• Batch DOI import — paste a list of DOIs, enrich all at once
• Smart filtering — journal, year, field tags, publication type, metadata completeness
• Progressive rendering — results appear per-source as they arrive, no waiting

• Summary — one-paragraph overview: what, how, why it matters
• Deep Analysis — 7 dimensions: contribution, motivation, methods, results, limitations, context, future work
• Multi-Paper Compare — systematic comparison across approaches, innovations, and research gaps
• Novelty Detection — identifies under-explored areas, contradictions, and promising directions

Connect to 300+ Chinese universities via CARSI. Access ScienceDirect, Scopus, JSTOR, and institutional databases without VPN, from anywhere in the world.

• D3.js force-directed graph — zoom, pan, drag, click to expand
• Keyword co-occurrence network — discover thematic clusters
• Author collaboration network — see co-authorship patterns
• Click any paper node to load its citing and referenced papers on-the-fly

• Multi-format export — RIS, BibTeX, CSV, EndNote XML
• Citation formatting — APA 7th, MLA 9th, GB/T 7714, Chicago 17th, Vancouver
• Zotero integration — SQLite direct read + MCP plugin + bi-directional tag sync + PDF full-text extraction
• Tag system — custom tags, color picker, Zotero sync
• Reading status — unread / reading / read tracking
• Batch operations — multi-select, one-click export
• Smart export — auto-organized sub-folders: PDF / RIS / BibTeX / CSV / Citations

Why not just use Google Scholar? Or ChatGPT? Or Perplexity?

git clone https://github.com/vanthree31/PaperLens.git
cd PaperLens
pip install -r requirements.txt
python main.py

The app opens in a native window (PyWebView) or your default browser.

Download from GitHub Releases, or build:

build.bat          # Windows → dist/PaperLens.exe

Open Settings in the app:

All AI processing runs fully offline with Ollama — no API key, no data leaves your machine.

1. Open Settings → select your university from the CARSI list
2. Log in with campus credentials
3. ScienceDirect, Scopus, JSTOR are now accessible from anywhere

Requirements: Python 3.9+, modern browser (Chrome/Edge recommended)

See ROADMAP.md.

Bug reports, feature ideas, and code contributions are all welcome.

Good first issues: good-first-issue

1. Fork the repo → 2. Create a branch → 3. Commit → 4. Open a PR

See CONTRIBUTING.md for details.

Feedback & Discussion (QQ Group)

QR Code	Group ID
	632743351

If PaperLens helps your research, consider supporting the project. ☕

Free for personal, academic, and research use.

Commercial use (including SaaS, paid services, or integrating into a for-profit product) requires prior written authorization. See LICENSE for details.

📧 For commercial licensing or partnership: vanthree31@gmail.com

PaperLens — Your AI Research Copilot
_{Find. Read. Understand. Discover.}