The integration of AI into educational tools is now a reality within everyone’s reach.
But behind the promises of AI lies a fundamental issue: cost.
The open-source and free assessment platform WASPL offers a built-in economic strategy for item generation.
Of course, you can use an API to generate your items: simply specify the topic and issue, choose an interaction type among those available, and indicate the number of items you need. If you have the budget, using an API provides a smooth and seamless user experience. The request is made from the application’s backend.
However, if you’ve decided to better control your spending after receiving your first AI service provider invoices, and you opt to invest in a local server capable of running recent AI models, you now have another option—more expensive upfront, certainly, but one with controlled and usage-independent costs.
Now let’s turn to the WASPL alternative, which, in addition to supporting remote or local APIs, allows you to generate a prompt that you can copy and paste into Claude, Mistral, or ChatGPT.
You retrieve the answer, check the JSON consistency (remove any trailing commas if necessary), and paste the generated code into WASPL.
The items will appear immediately; you’ll just need to distribute them across pages and adjust the scores as needed.
Here’s a short illustrative GIF:
