GHGitHub

Getting Started

Download, start and connect StoryBook AI on Windows, macOS or Linux.

StoryBook is open source and runs on your own computer. This guide walks you through the current setup so you can get StoryBook running and connect it to your local AI model.

1

Get StoryBook AI

Clone the repository with Git, or download it as a ZIP from GitHub and extract it to a folder on your computer.

$
git clone https://github.com/ocedo-apps/StoryBook-AI.git

StoryBook needs Node.js. The start scripts check whether Node.js is installed and point you to nodejs.org if it is missing.

2

Start StoryBook

Choose your operating system. The start scripts check for Node.js, run npm install the first time only, and then start StoryBook.

Windows

Double-click:

›
starta.bat

macOS

Double-click:

›
starta.command

On first launch, macOS may warn that the script is from an unidentified developer. Right-click starta.command, choose Open, then confirm. This should only be needed once.

Linux

Open a terminal in the StoryBook AI folder and run:

$
./starta.sh
Prefer to start it manually?
$
npm install
$
npm run dev

npm install contacts the npm registry while dependencies are installed. That setup step is separate from StoryBook's local-first runtime while you write.

3

Open StoryBook

Once StoryBook has started, open this address in your browser:

→
http://localhost:5175

The terminal will also show the local address when the development server starts.

4

Connect your local AI

Run a local model using Ollama, LM Studio or another compatible local endpoint. StoryBook defaults to a local address and is designed for AI running on your own computer or network.

Using Ollama? Allow StoryBook to connect.

StoryBook runs in your browser at http://localhost:5175. Ollama needs to allow requests from that address. Start Ollama with:

$
OLLAMA_ORIGINS=http://localhost:5175 ollama serve

If this is missing, StoryBook may appear unable to connect even though Ollama is running.

StoryBook lets you choose separate models for Writing and Review. A creative prose model can be used for drafting and rewriting, while a more analytical instruction-following model can handle review-oriented work. You can also use the same model for both.

Writing model

Used for Draft, Extend, Elaborate, Rewrite, Recast and other prose-oriented assistance.

Review model

Used for analytical work such as reviewing text and extracting structured information.

StoryBook AI model settings showing separate Writing and Review models and context window
Choose your local engine, Writing model, Review model and context window.
Context window: StoryBook lets you set how much context the model may use. With Ollama, Suggest from model can read the maximum context length reported by the connected model as a starting point.

There is no single best prose model for every writer or machine. Try models that fit your hardware, language and preferred style. StoryBook does not depend on one specific model.

5

Start writing

Create a manuscript, give it a title and open the first chapter. You can write directly in the editor or use planning tools such as Brainstorm, Synopsis and chapter briefs first.

You do not need to learn every feature before you begin. Start simple and add structure when the story needs it.

Interface: StoryBook is available in English, Swedish and Norwegian Bokmål, with light and dark themes. Your manuscript language is configured separately.

Explore StoryBook features →
6

Need help?

If StoryBook cannot see your AI model, first make sure the local AI server is running and that the configured endpoint matches the address shown by Ollama or LM Studio.

Ollama is running, but StoryBook still cannot connect?

Make sure Ollama was started with OLLAMA_ORIGINS=http://localhost:5175. Without that setting, the browser request can be blocked even though the Ollama server itself is running.

Open the project on GitHub →