aiCoder is a tool for building well structured javascript libraries.

Mike Molinari — Fri, 28 Feb 2025 07:04:35 +0000

Just wanted to share this project I have been working on.

It is an AI coding tool that takes naturel language prompts to generate code and automatically merge the new snippets in to your existing project files.

To merge the code in to an existing file we use an AST (abstract syntax tree) and surgically replace only the code blocks that need to be modified.

There is a live example of the merging technology demonstrated on this page. https://aicoderproject.com/

Github repo here.
https://github.com/mmiscool/aiCoder

If you want to try it out and you have an LLM api key just run the following commands.

cd project_directory
npx aicodertool

Open the browser and visit localhost:3000

Using ASTs to merge LLM generated snippets in to existing code files with surgical precision.

Mike Molinari — Mon, 23 Dec 2024 23:35:46 +0000

https://github.com/mmiscool/aiCoder

I’m thrilled to share aiCoder, a project that brings a fresh approach to AI-assisted JavaScript development! By harnessing the power of Abstract Syntax Trees (ASTs), aiCoder makes merging AI-generated code into existing projects smarter and more reliable than ever. Unlike traditional line-based methods that can lead to frustrating conflicts or overwrites, aiCoder works at the syntactic level, ensuring that every snippet integrates seamlessly while preserving the structure and logic of your code. It even keeps your comments intact, so your code stays clear and readable. I’m incredibly excited about how this tool is transforming the way we collaborate with AI, and I can’t wait to see how it boosts productivity and creativity for developers everywhere!

Slide deck explaining how the interesting bits work.
https://docs.google.com/presentation/d/1xdX09ELgW7lMU1E9KWIrpibUYVT1wdaiSvUhFhAT7EI/edit?usp=sharing

DEV Community: Mike Molinari

aiCoder is a tool for building well structured javascript libraries.

Using ASTs to merge LLM generated snippets in to existing code files with surgical precision.