π³ How to Reduce Credit Costs
Credit usage in Ludus AI is dynamic, varying based on the complexity of your requests, the length of conversation history, and the amount of project data processed. To maximize your efficiency, it is important to remember that Ludus is designed as an intelligent assistant to help Unreal Engine developers solve specific problems, not as a tool for automatically building entire games from scratch. By adopting a few smart habits, you can significantly extend the value of your monthly credits.
1. Choose the Right Modeβ
The most effective way to save credits is to use the appropriate mode for your task.
- Use Agent Lite (Faster & Cheaper): Ideal for general questions, "how-to" queries, definitions, Blueprint Generation, and 3D model generation. Agent Lite is faster and more cost-efficient than Agent. Both modes share the same context awareness within the plugin. Note: For complex C++ tasks, we recommend using Agent or MCP for best results.
- Use Agent (Better for Complex Tasks): Use it when you need superior reasoning for complex, multi-step tasks. Agent takes more time to process but delivers better results for intricate problems. Using Agent mode for a generic question like "What is an Actor?" might not be the most efficient use of your plan.
2. Keep Conversations Shortβ
Every time you send a new message, the AI re-reads the entire conversation history to maintain context.
- Start New Chats Frequently: As a chat gets longer, every simple reply costs more because it carries the weight of the entire history.
- The "One Topic, One Chat" Rule: Once you solve a specific problem (e.g., "Fixing the Character Movement"), close that thread and start a new one for the next task (e.g., "Creating a HUD"). This resets the context size to zero.
3. Leverage Free Tools: Ludus Insightsβ
Don't spend credits asking the AI to manually check your project for general optimization or code quality issues.
- Use Ludus Insights: The comprehensive project analysis report (accessible via the Plugin menu) is free. It scans your Blueprints, C++, and assets for performance bottlenecks and spaghetti code without consuming your interactive credits.
4. Efficient Promptingβ
- Be Specific: Vague prompts lead to follow-up questions. A 5-message exchange to clarify what you want costs 5x more than one well-crafted prompt.
- Batch Questions: If you have related questions, ask them in a single prompt rather than multiple short messages. This saves the overhead of processing the context multiple times.