From E-book to Mind Map: A Practical Guide to Turning Any Digital Book into a Visual Knowledge Graph

Three quick questions

  • •
    After finishing a 300-page technical book, do you only remember scattered ideas a week later?
  • •
    When taking notes, do linear highlights fail to show how chapters connect?
  • •
    Need to condense a long PDF report into a one-page mind map for your team—without drawing it by hand?

If you nodded at least once, this article gives you a zero-setup solution: drag an EPUB or PDF into a small open-source tool, grab a coffee, and come back to structured mind maps plus chapter-level summaries—all in plain English, no coding, no paywall.

1. What the tool actually does

Feature What you get When to use it
Text-Summary Mode ~250-word chapter abstracts + 500-800-word full-book recap Writing essays, quick reviews
Chapter Mind-Map Mode One interactive mind map per chapter Exam prep, slide decks
Whole-Book Mind-Map Mode Single overview map with drill-down nodes Executive briefings
Smart Cache Same file opens instantly on second visit Revising reports

In short: it automates “read thick, then thin”.

2. Five-minute walkthrough

All steps below are taken verbatim from the project README and tested on macOS, Windows, and Ubuntu.

2.1 Install (2 minutes)

# 1. Clone the repo
git clone <repository-url>
cd ebook-to-mindmap

# 2. Install dependencies
pnpm install        # recommended
# or
npm install

2.2 Start the dev server (1 minute)

pnpm dev

Your browser opens at http://localhost:5173.

2.3 Configure AI service (1 minute)

On first launch you’ll see a settings modal. Pick one provider:

Provider How to get a key Typical free quota
Google Gemini Google AI Studio → Create Key 60 requests/min
OpenAI GPT OpenAI Dashboard → API Keys Pay-as-you-go

Paste the key and keep the default model—no tuning needed.

2.4 Upload a book (10 seconds)

Click the big “Choose EPUB or PDF” button.
Accepted formats: .epub, .pdf, up to 50 MB.

2.5 Three quick settings (30 seconds)

Option Recommended value Why
Processing mode Start with “Text-Summary Mode” Fastest first impression
Book type Non-fiction for tech books Alters AI prompts
Smart chapter detection Off by default Most PDFs have a table of contents

Click Extract Chapters. A list appears on the right. Tick the chapters you want (all by default), then hit Process.

2.6 Read the results (1 minute)

When the progress bar hits 100 %, the left sidebar shows:

  • •
    Chapter summaries
  • •
    Chapter connection analysis
  • •
    Full-book recap

Copy-paste straight into your note app.

3. How it works (plain-English edition)

Here is the official architecture translated into everyday language.

3.1 End-to-end pipeline

graph TD
    A[Upload] --> B[Parse text]
    B --> C[Split into chapters]
    C --> D[AI summarizes each chapter]
    D --> E[AI links chapters]
    E --> F[Generate overview]
    F --> G[Cache locally]

3.2 How chapters are detected in a PDF

  1. Table of contents (first choice)
    Reads built-in bookmarks; accuracy ~99 %.

  2. Keyword heuristics (fallback)
    Looks for “Chapter 3”, “第 4 章”, etc. Good for scanned books.

  3. Fixed page chunks (last resort)
    Groups every 10 pages, ensuring ≥100 characters each.

“

Tip: If your scanned PDF fails step 1, toggle “Smart chapter detection” on.

3.3 How the AI writes summaries

Input snippet Example
Chapter title Chapter 5: Microservice Decomposition
Raw text 8 000 words
Prompt “Write 200-300 words in English covering: main points, key arguments, important concepts, chapter significance”

The AI returns a JSON blob; the UI renders it instantly.

3.4 Cache mechanics

  • •
    Cache key: filename + chapter hash
  • •
    TTL: 7 days
  • •
    Max entries: 100 (oldest evicted first)

Re-open the same book tomorrow and it loads in one second, no API cost.

4. Real-world playbooks

Playbook A: Finish “Computer Systems: A Programmer’s Perspective” in one week

Day Action Outcome
Mon Upload PDF → Text-Summary Mode 19 chapter abstracts
Wed Re-open cache → switch to “Chapter Mind-Map Mode” 19 interactive maps
Fri Re-open cache → switch to “Whole-Book Mind-Map Mode” 1 slide-ready overview

Playbook B: Turn a 200-page white paper into an executive deck

  1. Skip front matter – untick “Preface”, “Appendix”.
  2. Batch export – click “Export PNG” on the mind-map, drop straight into PowerPoint.
  3. Boss drill-down – click any node to reveal the exact source paragraph.

5. Frequently asked questions (FAQ)

Q1: Must I stay online?
A: First run needs the AI service; afterward cached results work offline.

Q2: Will the API break my wallet?
A: Gemini’s free tier grants 60 requests/min—ample for a 30-chapter book.

Q3: Encrypted PDFs?
A: Remove the password first; the tool cannot decrypt.

Q4: EPUB compatibility?
A: EPUB 2.0/3.0 supported. Kindle AZW3 needs conversion to EPUB first.

Q5: Can I switch to GPT-4?
A: Yes—in settings change base_url and model name to your OpenAI endpoint.

6. Power-user tweaks

Need How Benefit
Only core chapters Manually un-tick “Acknowledgements”, etc. ~30 % faster
Reduce map clutter Advanced → set “Sub-chapter depth” to 1 Cleaner visuals
Longer summaries Edit prompts/summary.txt, change 200-300 to 400-500 words Custom length
Shared workstation Settings → Clear cache per session Privacy

7. Developer quick-start for custom hacks

The stack is 100 % front-end; no server needed.

src/
├── services/pdfProcessor.ts   # chapter splitting logic
├── services/geminiService.ts  # AI prompts
├── services/cacheService.ts   # localStorage helpers

Example: To switch output language to Spanish, open prompts/summary.txt, replace the English prompt with Spanish instructions, restart pnpm dev, done.

8. Wrap-up: one diagram to rule them all

E-book file → Tool → [Text summary | Chapter map | Whole book map] → Copy/paste ready

Zero config, zero cost, zero code. Next time you face a thick book or lengthy report, let the tool “read thin” first, then dive deeper where it matters.

Happy mapping, and may your knowledge graphs always stay clear.