I’m working on a Claude skill that generates Angular components for a big enterprise project (NX monorepo, porting from a legacy codebase). The main goal of my work is simple, help less experienced devs move faster by handing them a draft component that already follows our project standards: naming, structure, etc…
Somewhere along the way I started playing with Graphify, a library that builds a graph of your project so your AI agent can find its way around without falling back on Grep and burning tokens.
After a bit of testing, I decided to keep it. Not really for the token savings (in my opinion those are minor than expected), but for another reason. Which one? Read the article and find out 😜
🇮🇹 (Versione italiana) <a href="https://www.thinkasadev.com/graphify-knowledge-graph-per-claude-code/" target="_blank" rel="noopener nofollow">https://www.thinkasadev.com/graphify-knowledge-graph-per-claude-code/</a>
🇬🇧 (English version) <a href="https://www.thinkasadev.com/en/graphify-knowledge-graph-for-claude-code/" target="_blank" rel="noopener nofollow">https://www.thinkasadev.com/en/graphify-knowledge-graph-for-claude-code/</a>
<a href="#claude" target="_blank" rel="noopener nofollow">#claude</a> <a href="#claudecode" target="_blank" rel="noopener nofollow">#claudecode</a> <a href="#skill" target="_blank" rel="noopener nofollow">#skill</a> <a href="#graphify" target="_blank" rel="noopener nofollow">#graphify</a> <a href="#graphifyy" target="_blank" rel="noopener nofollow">#graphifyy</a> <a href="#token" target="_blank" rel="noopener nofollow">#token</a> <a href="#angular" target="_blank" rel="noopener nofollow">#angular</a> <a href="#monorepo" target="_blank" rel="noopener nofollow">#monorepo</a> <a href="#ai" target="_blank" rel="noopener nofollow">#ai</a> <a href="#IA" target="_blank" rel="noopener nofollow">#IA</a>

Giorgio

A developer shares their experience integrating Graphify with Claude Code to generate Angular components in an NX monorepo enterprise project. The goal is to help junior developers produce draft components that follow project standards. While token savings were less significant than expected, the author found another compelling reason to keep Graphify — building a knowledge graph so the AI agent can navigate the codebase more intelligently without relying on grep-based searches.

Graphify and Claude Code: save tokens and get better answers