Upgrading a Gaming AI: Exciting Experiments in Data Collection and 検索拡張生成 (RAG)
Infrastructure#rag📝 Blog|Analyzed: Apr 18, 2026 14:16•
Published: Apr 18, 2026 12:00
•1 min read
•Zenn LLMAnalysis
This is a fantastic hands-on exploration of building a specialized AI assistant using 検索拡張生成 (RAG) and web scraping techniques. The author provides a highly practical guide on expanding a Large Language Model (LLM)'s knowledge base by successfully gathering rich strategy data from mobalytics.gg. It also brilliantly highlights the importance of digital ethics, showing a commendable commitment to respecting terms of service and robots.txt rules during data collection.
Key Takeaways
- •Successfully gathered rich strategy data for Slay the Spire 2 using Python's requests library from mobalytics.gg.
- •Encountered API access restrictions with Reddit, highlighting the growing challenges of data sourcing in 2026.
- •Demonstrated ethical AI development by voluntarily abandoning scraped data from a domestic site due to potential commercial terms of service conflicts.
Reference / Citation
View Original"I wanted to try upgrading it with more data this time. Conclusion: The data went in. But 検索拡張生成 (RAG) couldn't find it."