Adventures in MuleSoft + AI
Last updated
Last updated
Curious if AI can help you build MuleSoft applications faster and cleaner? In this dev experiment, I test out Cursor AI + Anypoint Code Builder to see if it's up to the task.
In this follow-up to my previous video, I took the MuleSoft best practice template I generated earlier and asked Cursor AI to build a new Mule app using the same structure and standards.
๐ง Can you teach Cursor AI to follow your MuleSoft best practices automatically?
In this video, I explore exactly that โ using an .mdc file to define custom rules that help Cursor AI generate better MuleSoft code, with a cleaner structure and reusable components.
After struggling in Part 2, this time I leveraged Cursor rules to guide my MuleSoft app creation. Itโs still not perfect, but if youโre specific enough, Cursor becomes a powerful AI coding assistant โ even for integration workflows.
In this video, I test CurieTech AI's DataWeave code generator to see how well it handles real-world MuleSoft transformations.
Can AI really generate the most performant DataWeave code? In this video, I put CurieTech AIโs DataWeave code generator to the test to see if it could choose the fastest transformation logic โ using nothing but sample input and expected output.
In this video, I put two AI coding tools head-to-head to find out which one can generate MUnit tests for MuleSoft more easily, accurately, and efficiently.
In this video, I take the same MuleSoft best practices rules file I used in a previous test with Cursor AI, and this time I run it through CurieTech AI's Code Enhancer Agent. The results? Not only did Curie generate clean, best-practice-compliant code... it also opened a pull request directly in my GitHub repo. ๐คฏ
In this video, I dive into the Code Insights agents available in CurieTech AI โ specifically Single Repo Code Lens, Code Review Lens, and Multi Repo Code Lens.
In this video, I test CurieTech AIโs new API Spec Generator agent โ a powerful tool that helps developers instantly generate OpenAPI (OAS) and RAML specifications from prompts, existing specs, or even vague ideas.