GitBook-Forge

A high-performance CLI forged in Rust to convert any GitBook site into clean, AI-ready Markdown.

It discovers all pages from a site's table of contents, renders them with a real browser, and saves each page as a separate .md file—creating a perfect corpus for your RAG pipelines and LLM fine-tuning.

---

Key Features

GitBook-Forge is built for AI developers who need to ingest documentation into RAG pipelines or for model training.

• AI-Ready Markdown: Intelligently converts complex elements like tables and code blocks into clean Markdown while skipping non-textual assets like images and videos.
• Blazing Fast: Built in Rust with a concurrent architecture to download and process pages in parallel.
• Handles Private Sites: Natively supports authentication for private GitBooks.
• Zero Dependencies: Ships as a single, self-contained executable with no need for Node.js or Python.

---

Getting Started

Method 1: Download a Pre-compiled Binary

1. Download: Go to the Releases page.

2. Extract: Download the archive for your OS (Linux, macOS, Windows) and extract it.

3. Run: The archive contains a single executable. Run it from your terminal.

   (On Linux/macOS, you may need to make it executable: chmod +x ./gitbook_forge)

Method 2: Build from Source

Requires the Rust toolchain.

1. Clone & Build:

   git clone https://github.com/Ebrahim-hamdy/GitBook-Forge.git
   cd GitBook-Forge
   cargo build --release

2. Run: The optimized executable is located at target/release/gitbook_forge.

---

Usage

The basic command is gitbook_forge <URL> [OPTIONS].

Quick Example

# Download an entire site to a specific directory
gitbook_forge "https://docs.gitbook.com/" -o ./gitbook-docs

Command Reference

Flag (Short/Long)	Argument	Description	Default
(none)	<URL>	The root URL of the GitBook site to download.	(Required)
-o, --output	<PATH>	The directory to save the Markdown files.	./output
-s, --single-page	(none)	Download only the single page specified by the URL.	false
--no-images	(none)	Skip downloading all image assets.	false
--auth	(none)	Enable authentication for private sites.	false
-u, --username	<EMAIL>	Username (email) for authentication.	none
-p, --password	<PASSWORD>	Password for authentication (prompted if omitted).	none
--help	(none)	Show detailed help information.

License

This project is licensed under the MIT License. See the LICENSE file for details.