Reduced errors using the AI Builder

With this week's updates, we've brought two new improvements that change the way how the agent works on your requests.

Up until last week, the code quality check tool commands (which help reduce errors) were not AI friendly in their responses leading to wasted context and token usage.

This has since been optimised, leading to 20-50% savings in credit usage for each code quality check.

Instead of the agent needing to read entire files to find the code it’s looking for, it can search for snippets of code and text within files across the entire project.

This improves accuracy, reduces context/memory usage in a chat, and therefore reduces credit spend. Another minor benefit is that it also allows for more time before you need to start a new chat by using less memory.

From now on, the AI builder will require lesser credits (and effort) to work on your design prompts.

If you want to test it out and don't have access to the AI Builder yet, sign up for the waitlist here to get your hands on it.