The problem TalkToCode solves

• When I joined an organization as a new developer, my first task was to go through hundreds and thousands of lines of cluttered code lacking proper documentation. This made it incredibly time-consuming and challenging for me to understand and modify the code.
• TalkToCode acts like a senior developer, helping you understand large codebases easily. You can ask any doubts and queries you have about the codebase, and not only that, but it also generates documentation for you and your teammates :D
• Understanding code is a significant issue, with developers spending 60% of their time reading and comprehending it. Reading large, complex, and undocumented codebases, as well as other developers' code, can be hectic, time-consuming, and hinder productivity.
• When faced with doubts about code, such as where the Razorpay API is called and how it works, or which functions are linked to it, in-depth explanations are crucial.
• Knowing which file resides where, under which directory, or which API is called in which file, helps navigate through the codebase.
• TalkToCode generates automatic AI-generated code documentation for the entire codebase, providing clarity and streamlining development workflows.

Challenges I ran into

• To scrape large codebases from repositories via the GitHub API, given the GitHub API's limitations on free usage and scraping capabilities.
• Sending the full codebase to the Gemini API for AI-generated documentation. With large codebases, the input tokens and parameters increase accordingly.
• Enhancing TalkToCode to support multiple technologies, including Python, Flutter, etc.