🧠 Devlogue — A Journaling RPG

Devlogue is a journaling RPG web app that gamifies self-reflection and knowledge exploration. As you write about your day, debug your thoughts, or explore ideas, a game world grows around you — with quests, NPCs (not implemented), a world map(not implemented), and your own evolving avatar(not implemented).

✨ Built for journaling enthusiasts, and RPG lovers — Devlogue turns your thoughts into adventures.

---

🧩 Key Features

• 📝 Rich Text Journaling using Quill.js with autosave.
• 🤖 AI-powered Journal Companion (Gemini AI) to generate:
  • Reflective prompts
  • Summary of past entries
  • Quests derived from writing patterns
• 🧍‍♂️ Animated Rive Avatar that reacts to entries (not implemented) and clicks.

---

🚀 Installation Guide

You’ll need:

• Node.js v18+
• Git (optional but helpful)
• A modern browser (Chrome, Firefox, Edge)

🖥️ For all platforms (Linux, macOS, Windows):

Devlogue

A journaling app with AI-powered reflection quests and themes.

Getting Started

1. Clone the Repository

git clone https://github.com/yourusername/devlogue.git
cd devlogue

2. Install Dependencies

npm install

3. Set Up Ollama for AI Integration

This app uses Ollama to run local LLMs for journal analysis.

a. Install Ollama

Follow the instructions at ollama.com/download to install Ollama for your platform.

b. Pull a Model

You can use deepseek-r1 (recommended) or any other supported model.
To pull deepseek-r1, run:

ollama pull deepseek-r1

Or, for another model (e.g., llama3):

ollama pull llama3

c. Start Ollama

Ollama usually runs automatically after installation. If not, start it with:

ollama serve

d. Configure the Model in Code

By default, the app uses deepseek-r1.
If you want to use a different model, open
src/routes/+page.svelte
and find this section in the analyzeEntry function:

body: JSON.stringify({
    model: "deepseek-r1", // or another model you have pulled
    prompt,
    stream: false,
}),

Change "deepseek-r1" to your preferred model name (e.g., "llama3").

---

4. Run the App

npm run dev

Visit http://localhost:5173 in your browser.

---

Notes

• Make sure Ollama is running and the model you specify is pulled.
• The AI features will not work if Ollama is not running or the model is not available.
• For best results, use a model that supports instruction-following and JSON output.

---

License

Apache 2.0

Credits

Rive Avatar animation by mtdsgnco