⏳ This skill is pending AI review.
Scores will appear once the review pipeline completes.
paper-to-slides
End-to-end academic paper analysis and presentation generation. Use when a user provides a PDF paper (local path or arXiv URL) and wants (1) a deep research report (dual-mode Part A + Part B), (2) an HTML slide deck for group meeting / seminar / conference talk, or (3) both. Triggers on "read this paper and make slides", "parse this paper into a presentation", "论文研读+做PPT", "解析论文做成slides", or any request combining paper analysis with presentation output. Supports arXiv links (abs/pdf), DOI URLs, and direct PDF URLs.
// RATINGS
// README
English | 中文
Paper to Slides
An AI skill that deep-reads academic papers and turns them into polished HTML presentations — give it a PDF or arXiv link, get back a structured research report and a ready-to-present slide deck.
Philosophy: Reading a paper should produce reusable artifacts, not just understanding. A single command gives you both a deep analysis and presentation-ready output.
What You Get
Give the skill a paper (local PDF, arXiv link, or any PDF URL), and it produces:
- Research Report — A dual-mode deep-read: Part A is a full academic analysis (structured abstract → methodology → results → discussion), Part B distills the core logic chain into a one-line summary
- HTML Slide Deck — A zero-dependency, animated presentation with 10 style presets, browser-based inline editing, and responsive design that works on any screen
You can request either output alone or both together.
Quick Start
- Install the skill in your agent (OpenClaw or Claude Code)
- Say "read this paper and make slides" with a PDF path or arXiv link
- The agent handles everything — extraction, analysis, style selection, generation
解析 https://arxiv.org/abs/2603.02096，组会汇报，亮色，15页左右
Deep-read /path/to/paper.pdf, generate report and slides, dark theme
The agent will ask about slide preferences (length, theme, editing) if not specified. Or provide them upfront to skip the questions.
Style Gallery
10 built-in style presets. Open live preview →
Light Themes
| Swiss Modern ⭐ | Notebook Tabs | Pastel Geometry |
|---|---|---|
 |  |  |
| Clean, precise, academic | Editorial, organized, tactile | Friendly, approachable |
| Split Pastel | Vintage Editorial | Paper & Ink |
|---|---|---|
 |  |  |
| Playful, creative, dual-tone | Witty, confident, retro | Literary, thoughtful |
Dark Themes
| Bold Signal | Neon Cyber | Dark Botanical | Terminal Green |
|---|---|---|---|
 |  |  |  |
| High-impact, bold | Futuristic, techy | Elegant, artistic | Hacker aesthetic |
Installation
OpenClaw
# From ClawHub (coming soon)
clawhub install paper-to-slides
# Or manually
git clone https://github.com/zhangguanghao523/paper-to-slides.git ~/.openclaw/skills/paper-to-slides
Claude Code
git clone https://github.com/zhangguanghao523/paper-to-slides.git ~/.claude/skills/paper-to-slides
System Dependency
The skill needs pdftotext for PDF text extraction:
# macOS
brew install poppler
# Ubuntu/Debian
sudo apt install poppler-utils
That's it. No Python packages, no npm, no build tools. Generated slides are pure HTML with inline CSS/JS.
Architecture
This skill uses progressive disclosure — the main SKILL.md is a concise workflow map, with supporting files loaded on-demand only when needed:
| File | Purpose | Loaded When |
|---|---|---|
SKILL.md | Core two-phase workflow | Always (skill invocation) |
prompts/part-a-template.md | Part A deep academic report template | Phase 1 (report writing) |
prompts/part-b-template.md | Part B core logic extraction template | Phase 1 (report writing) |
prompts/slide-styles.md | 10 curated visual presets | Phase 2 (style selection) |
prompts/slide-template.md | HTML architecture, viewport CSS, JS controller | Phase 2 (generation) |
Built On
| Component | Role |
|---|---|
| paper-parse | Dual-mode report templates (Part A + Part B) |
| frontend-slides | Zero-dependency HTML presentation framework |
| poppler | PDF text & figure extraction (pdftotext / pdftoppm) |
Requirements
- An AI agent (OpenClaw, Claude Code, or similar)
pdftotext(from poppler) — for PDF text extraction- Internet connection (for arXiv downloads and Google Fonts in generated slides)
Credits
Built with OpenClaw. Presentation engine adapted from frontend-slides by @zarazhangrui. Report templates adapted from paper-parse.
License
MIT — Use it, modify it, share it.
// HOW IT'S BUILT
KEY FILES