Keyforge

Type

Webapp

KeyForge is a new card game released by Fantasy Flight Games. 10Clouds built an official mobile and web application for it. With Master Vault, users can quickly scan their cards, add notes or track their stats.
In just five days after its release, players have registered 100,000 decks online, and the app has been downloaded more than 10,000 times. The tool was warmly received by the community.

Services

Web
Design

Our challenge

The KeyForge game platform is used by thousands of players all over the world every day. One of the main functionalities on the website is card decks’ registration which allows players to add their deck of cards to the collection. So far almost 1,8 million unique decks have been registered on the platform and all users can see the details of each one on the platform. Having 1,7M unique decks listed on 17,000 pages is a huge challenge for the infrastructure and efficiency.

How we made it happen

We started by analysing the current situation and understanding what causes the poor performance of the page which lists all the registered decks.

We quickly found that the problem is not only on the code and infrastructure level but also in the way users were interacting with the page. It turned out that there were a number of unofficial pages which were scraping the data from the site and causing performance issues.

Our development team implemented multiple solutions to optimise the database and page performance - indexing, queries optimisation, caching etc. For the issue with data scrapers we decided not to fully block their activity but to control it by introducing 3 level traffic throttling. All of the implemented improvements resulted in a much smoother experience when using the KeyForge website.

Check other case studies

INTERNAL PRODUCT, AI INTEGRATION, CHATBOT

10Clouds AI Assistant

10Clouds Assistant is a powerful tool that can collect data from company handbooks, employee information, project details, and more to provide instant feedback. The process of creating this tool involved several stages, such as downloading data from the Confluence API and saving it to a PostgreSQL database. Next, we cleared the data of HTML tags and added it to Pinecone using Llama-index, a library for importing data. Our AI tool can quickly and accurately provide users with the information they need, making it an invaluable asset for any organization.

See more

Get the full story on how we help companies get results faster.

Contact us