Rogue Adventure Assistant

🤖 RA Assistant

Enter your OpenAI API key to get started

OpenAI API Key

Model

Knowledge Base

Cards

Skills

Enemies

Embeddings

Settings

Context Cards (retrieved per query)

💬 Chat Sessions

Tips

• Ask about card synergies, deck builds, or strategies

• Mention your class for tailored advice

• The assistant knows all card abilities and stats

📖 User Guide & Instructions

Getting Started:

1. Enter your OpenAI API key above

2. Click "Initialize Assistant" to load the knowledge base

3. Start asking questions about Rogue Adventure!

API Key Security:

• Your key is stored locally in your browser only

• Never shared with anyone except OpenAI's API

• Get a key at platform.openai.com

Example Questions:

• "What are the best cards for Demon builds?"

• "How does Poison synergize with Assassin?"

• "Compare Shield vs Resistance mechanics"

• "What's a good starter deck for Wizard?"

Model Selection:

• Budget: Great for quick questions

• Balanced: Better reasoning

• Premium/Expensive: Complex analysis (costs more)

Best Demon synergy cards? How does Poison work? Explain Shield vs Resistance Good cards for Wizard? What beats high-HP enemies?

Welcome to the Rogue Adventure Assistant! Enter your API key and click Initialize to start.

Conversation: 0 tokens

Input: 0 tokens

Total: 0 / 128k

🤖 Model Settings

Model

GPT-4o Mini

$0.15 / $0.60 per 1M tokens (input/output)

Fast, affordable for most tasks. Great for quick questions.

0.7

💬 Conversation Settings

Set to 0 for unlimited (token-based limit only)

📝 System Prompt

Customize how the assistant behaves. This is sent at the start of every conversation.

⚠️ Advanced Settings

1.0

0.0

Enable Debug Mode

Shows token counts and API details in console

🔐 Developer Tools

Enter password to unlock developer features:

🎉 Congrats, you found the hidden Regenerate Embeddings button!

This is just used to regenerate the database used for the AI. It's purely here for me to update the bot when I add new stuff.

All it will do for you is waste your money, so don't mess with it lol.
Or do, I don't care :P

🔄 Reset Options

🚀 Getting Started

Get an API Key: Visit platform.openai.com to create your OpenAI API key
Enter Your Key: Paste it in the "OpenAI API Key" field in the left panel
Initialize: Click "🚀 Initialize Assistant" to load the game database
Start Chatting: Ask questions about Rogue Adventure!

❓ Frequently Asked Questions

Is my API key secure?

Yes! Your API key is stored only in your browser's localStorage and is never sent anywhere except directly to OpenAI's API. We don't have any backend servers - everything runs in your browser.

How much does it cost?

Costs depend on the model you choose. GPT-4o Mini is the cheapest at ~$0.15 per million input tokens. A typical conversation might use 2,000-5,000 tokens, costing less than $0.01. Premium models like GPT-5.1 cost more but provide better reasoning.

What does the assistant know?

The assistant has knowledge of all 391 hero cards, 186 skills, 331 enemies, and their stats, abilities, and synergies from Rogue Adventure v3.9.3. It uses this knowledge to answer your questions about strategies, deck building, and game mechanics.

How does the retrieval system work?

The assistant uses a multi-layer retrieval system to find relevant game data:

📝 Lexical Search - Finds exact name matches and keyword matches
🏷️ Structured Search - Filters by race, class, type, rarity, and mechanics
🧠 Semantic Search - Uses AI embeddings to find conceptually similar content

Results are combined and weighted to give you the most relevant information. Expand "📚 Sources" on any response to see what was found and how.

Why are some responses cut off?

This can happen if the "Max Output Tokens" setting is too low. Go to Settings and increase it (default is 1500, try 3000-4000 for longer responses). Reasoning models (o3, o4, gpt-5) use tokens for "thinking" so need higher limits.

What's the difference between models?

Budget models (GPT-4o Mini, Nano variants) are fast and cheap, good for simple questions.
Balanced models (GPT-4o, Mini variants) offer good quality/cost ratio.
Premium models (GPT-5, GPT-5.1) provide the best reasoning but cost more.
Reasoning models (o3, o4) excel at complex analysis but are slowest and most expensive.

How do I edit or regenerate messages?

Edit any message: Hover over a message and click ✏️ Edit. You can:
• Save - Edits the message in place, conversation continues unchanged
• Save & Regenerate - (User messages only) Creates a new branch and gets a new AI response
• Cancel - Discard changes

Regenerate AI response: Hover over any assistant message and click 🔄 to generate an alternative response.

How does conversation branching work?

When you regenerate a response or use Save & Regenerate, a new branch is created. You'll see navigation arrows (‹ 1/2 ›) that let you switch between different versions. All versions are saved, and you can switch between them anytime.

How do I manage chat sessions?

Use the 💬 Chat Sessions panel in the sidebar:
• + New - Start a new conversation (current is auto-saved)
• 💾 Save - Manually save current session
• Click any session to load it
• ✏️ Rename or 🗑️ Delete sessions

Sessions are automatically named from your first message and stored in your browser.

💡 Example Questions

"What are the best cards for a Demon build?"
"How does Poison synergize with Assassin class?"
"Compare Shield vs Resistance - when should I use each?"
"What's a good starter deck for Wizard in Hell mode?"
"Which cards counter high-HP enemies?"
"Explain how combo mechanics work"
"What are the strongest legendary cards?"

💡 Tip: You can use casual terms - the assistant understands synonyms like "block" → Shield, "draw" → Insight, "devils" → Demon, "mage" → Wizard, etc.

⌨️ Keyboard Shortcuts

Enter Send message Shift + Enter New line in input

💰 Model Pricing Reference

Prices are per 1 million tokens (input/output)

💰 Budget-Friendly

GPT-4o Mini: $0.15/$0.60 • GPT-4.1 Nano: $0.10/$0.40 • GPT-5 Nano: $0.05/$0.40

⚖️ Balanced

GPT-4o: $2.50/$10 • GPT-4.1 Mini: $0.40/$1.60 • GPT-5 Mini: $0.25/$2 • o4-mini: $1.10/$4.40

⚠️ Premium

GPT-4.1: $2/$8 • GPT-5/5.1: $1.25/$10 • o3-mini: $1.10/$4.40

🔴 Expensive

o3: $2/$8 • o4-mini Deep Research: $2/$8

ℹ️ About

Rogue Adventure Assistant is a RAG (Retrieval-Augmented Generation) chatbot built for the Rogue Adventure data mining project. It combines game data with LLM capabilities to provide accurate, context-aware responses about the game.

Version 2.2 • Game Data: v3.9.3 • Multi-layer RAG