COGNOSCERE BookSummarizer v2

Sistem otomatis untuk mengekstrak dan merangkum koleksi ebook PDF menjadi review komprehensif 8.000–15.000+ kata menggunakan dual AI engine.

Dual Engine

Engine	Kecepatan	Cara Kerja
Kimi API	~10 menit/buku (sequential)	API langsung ke Kimi K2.5
Claude Web	~3 menit/buku (4-5 parallel)	Browser automation ke Claude.ai

Fitur

• Aplikasi macOS — Klik dua kali BookSummarizer.app, browser terbuka otomatis
• Dual Engine — Pilih Kimi API (stabil) atau Claude Web (cepat, parallel)
• 4-5 Tab Parallel — Claude mode membuka 4-5 tab browser secara bersamaan
• Auto-Kategorisasi — 13 kategori otomatis (Finance, Psychology, Parenting, dll)
• Folder Terstruktur — Output di output/Completed/<Kategori>/<Judul Buku>/
• COGNOSCERE v2.0 — Review mendalam 9 section, bukan rangkuman biasa
• Dual Output — Markdown + DOCX profesional per buku
• Resume Otomatis — Kalau berhenti, lanjut dari buku terakhir
• Zero CLI — Semua lewat antarmuka web

Cara Menggunakan

1. Clone & Setup

git clone https://github.com/subkhanibnuaji/BookSummarizer.git
cd BookSummarizer
python3 setup_v2.py

2. Build App (sekali saja)

python3 build_app.py

3. Jalankan

Klik dua kali BookSummarizer.app di Finder.

Browser otomatis terbuka ke http://localhost:5000.

4. Pilih Engine

Di halaman Settings:

• Masukkan API Key Kimi (untuk Kimi mode)
• Pilih engine: Kimi API atau Claude Web
• Pilih folder ebook PDF

5. Proses

Buka Library → centang buku → Start Processing.

Struktur Output

output/Completed/
├── Finance & Investing/
│   ├── The Art of Spending Money/
│   │   ├── COGNOSCERE - The Art of Spending Money.docx
│   │   ├── COGNOSCERE - The Art of Spending Money.md
│   │   └── The Art of Spending Money.pdf
│   └── ...
├── Technology & AI/
│   └── Building Applications with AI Agents/
│       └── ...
├── Psychology & Mind/
├── Self-Development/
├── Relationships & Social/
├── Parenting & Family/
├── Health & Wellness/
├── Business & Innovation/
└── General/

COGNOSCERE Review

Section	Isi
Metadata & Core Thesis	Info buku, tesis utama, signifikansi
Structural Map	Peta visual seluruh bab
Master Concept Inventory	Semua framework dan model
Chapter Deep Dive	Rangkuman per bab
Glossary	Istilah penting
Integration Map	Koneksi antar konsep
Application Guide	Roadmap implementasi
Mastery Verification	Pertanyaan pemahaman
Further Reading	Buku terkait

Bahasa: Bahasa Indonesia dengan istilah teknis English.

Persyaratan

• macOS 12+
• Python 3.8+ dari python.org atau Homebrew
• Koneksi Internet
• API Key Kimi (untuk Kimi mode) — gratis di kimi.com/code/console
• Login Claude.ai (untuk Claude Web mode)

Teknologi

Komponen	Teknologi
AI Engine 1	Kimi K2.5 via API
AI Engine 2	Claude.ai via Playwright
PDF Extraction	pdfplumber + pypdf
Document	python-docx
Web UI	Flask + vanilla JS
Browser Automation	Playwright (Chromium)
Metodologi	COGNOSCERE v2.0

Project Structure

BookSummarizer/
├── BookSummarizer.app      ← macOS app (klik dua kali!)
├── build_app.py            ← Build/rebuild .app
├── app.py                  ← Flask web app
├── engines/                ← AI engines
│   ├── kimi_engine.py      ← Kimi K2.5 API
│   ├── claude_web_engine.py ← Claude.ai browser automation
│   └── claude_prompts.py   ← COGNOSCERE prompts
├── services/               ← Business logic
│   ├── categorizer.py      ← Auto-kategorisasi buku
│   └── project_mapper.py   ← Claude project discovery
├── summarizer.py           ← Core Kimi summarizer
├── pdf_extractor.py        ← PDF text extraction
├── docx_creator.py         ← DOCX generator
├── setup_v2.py             ← Setup + Playwright install
├── start.py                ← Auto-launcher
└── requirements.txt        ← Dependencies

Author

Subkhan Ibnu Aji

License

MIT License

---

<p align="center"> <strong>COGNOSCERE v2.0</strong> — Dual Engine Knowledge Extraction<br> Kimi K2.5 + Claude.ai · macOS Edition </p>