Onboard-CLI is a local-first development tool that leverages AST and large language models to visualize the structure of massive, complex software and automatically block bad code before it is committed.
Imagine entering a massive library filled with tens of thousands of books. Every shelf is packed, but you have no idea which book is where or how they are connected. The helplessness a developer feels when joining a new project or dealing with a large-scale software system is quite similar.
Recently introduced on the developer community Hacker News, Onboard-CLI is a new tool designed to solve this feeling of helplessness. It serves as a “compass” to help you find your way through the maze of complex code.
Why is it getting attention?
Modern software is becoming increasingly massive and complex in structure. Developers spend a tremendous amount of time trying to figure out which feature is connected to what among tens of thousands of files. Maintenance, in particular, becomes a painful ordeal when “spaghetti code”—where features are intricately tangled together—is involved.
Onboard-CLI goes beyond just reading code; it visualizes the overall structure and prevents bad coding habits from infiltrating the project. When a developer wonders, “Is it safe to modify this code?”, it provides an immediate structural view, maximizing productivity and preventing unexpected accidents.
Easy Understanding: The AI Librarian That Maps Structure
Onboard-CLI uses two core technologies to organize complex code.
The first is AST (Abstract Syntax Tree) parsing. Simply put, when the computer reads code, it doesn’t just see raw text; it breaks down the grammatical meaning and connection structure into a tree-like map, much like analyzing the structure of a sentence [Source 2, Source 5]. Metaphorically, this is like using a filter in a smartphone photo app to clearly separate the elements of a picture.
The second is LLM (Large Language Model). Based on the parsed code information, this model helps developers gain a deeper understanding of the code [Source 2].
The analyzed code is then rendered into an intuitive map using a tool called ‘React Flow canvas.’ You can grasp the flow of the code at a glance, much like looking at a subway map [Source 5].
Current Status: A Fast Local Analyst
For security and privacy, Onboard-CLI takes a local-first approach, running directly on the developer’s machine [Source 6]. What is even more surprising is its processing speed. It has optimized concurrency design to the extreme, enabling it to analyze over 50,000 files in under 10 seconds [Source 4].
Furthermore, if a developer accidentally attempts to add poor dependencies or write spaghetti code, it automatically blocks this locally before the commit (save) is finalized [Source 4]. It is like having a navigation system that immediately warns you, “That’s a dead end!” if you take a wrong turn while coding. The tool is currently open-source and available for anyone to use via GitHub [Source 1, Source 2].
Future Outlook
Tools like Onboard-CLI are likely to become standard “basic skills” for developers. The core competence of a developer is shifting from simply writing good code to how quickly one can understand the entire code structure and keep it maintainable. The creator is currently operating a beta version and refining features based on feedback from engineers [Source 6]. As AI analysis technology becomes more sophisticated, an era will come where even novice developers can instantly understand and manage the structure of massive systems.
MindTickleBytes’ AI Reporter Perspective
The essence of coding is transitioning from “feature implementation” to “complexity management.” Onboard-CLI proves that AI can be usefully employed to draw the grand map of software architecture beyond simple code auto-completion. This trend of developers visually understanding code and preemptively preventing bad patterns will play a major role in creating a healthier and more robust software ecosystem.
References
- Show HN: Onboard-CLI, an AST-based tool to detect …
- Developer launches Onboard-CLI, an LLM-powered and AST …
- Show HN: Onboard-CLI, a LLM powered and AST-based tool to visualize codebase
- Show HN: Onboard-CLI, an AST-based tool to detect …
- Show HN: Onboard-CLI, an AST and LLM-based codebase visualization tool…
- @markproduct I built Onboard-CLI a local-first, AST-powered …
- Better HN - bhn.vercel.app
-
[Onboard-CLI: Visualizing Complex Code Architecture and Boundary Protection Zeli](https://zeli.app/zh/story/48836813)
- Image recognition
- AST (Abstract Syntax Tree) parsing
- Simple text search
- Analyzes over 50,000 files in under 10 seconds
- Takes over an hour
- Relies solely on cloud servers
- Manual human review
- Automatically blocks bad code and dependencies before commit
- Does not delete code