Google working on making Gemini able to analyze entire folders with code files

0comments
Gemini's logo next to a phone showing the app's home screen on a dark background.
Gemini, the generative AI by Google, will be getting a notable upgrade that will allow it to analyze entire code folders. Right now, the feature has been spotted on the web and within the Android app and seems to be gradually rolling out.

Generative AI has quickly become the hottest topic in the tech industry. In the last couple of years, it's been gaining features and abilities and has steadily been becoming a very useful tool for people. One of the things generative AI can help with is analyzing your code and suggesting improvements, even debugging.

Code analysis often requires looking at many interconnected files. Of course, this isn't always possible, especially with current-generation AI assistants. However, Google is working on changing that with its assistant, Gemini, and is giving it the ability to analyze a codebase with multiple files in it.

The folks at Android Authority have discovered code hidden in the Google app v15.47.28 beta that indicates Google is working on the feature for Gemini to analyze code folders.

Right now, you can upload files to Gemini for analysis, and this can be used for code files. But uploading an entire folder is not possible at the moment. ChatGPT, Gemini's rival, already supports multi-file folders, and Google is now working to catch up.

TestingCatalog spotted these changes on Gemini on the web. You seem to be able to select one folder per chat. The limit there is up to 1,000 files per folder, with a total size of 100 MB.


Once the upload is completed, you can ask Gemini some specific questions about the code and the files.

Right now, it's not clear when a more global rollout would be put in place. What's also not clear at the moment is whether this feature will be exclusive to paid subscribers.  

Although I don't code, I can see how this feature will be helpful for developers and save them time from digging into multiple files. I really like how AI can learn to do annoying things while we, humans, can focus on what's more important: including creating unique code for exciting programs to begin with.

Recommended Stories

Loading Comments...
FCC OKs Cingular\'s purchase of AT&T Wireless