Look, the AI tooling space has exploded. There are hundreds of tools out there, and most of them are noise. After trying a bunch of them, I narrowed it down to the ones I actually use and genuinely recommend — tools that fit into a real engineering workflow and make a meaningful difference.
Let's get into it.
1. Cursor — Your Entry Point to Agentic AI
If you're just starting to explore AI-assisted development, start here. Cursor is the gateway drug to working with LLMs in your day-to-day coding life. It helps you get comfortable with what AI can actually do — autocomplete, refactoring, explaining code, writing tests — before you go deeper into more powerful (and complex) tools.
Think of it as your training wheels, but in the best way possible.
2. Search & Business Analysis: Perplexity + Gemini Pro Deep Research
For research tasks — competitive analysis, digging into a new tech stack, understanding a domain — these two are my go-to combo:
- Perplexity is great for quick, sourced answers. It's like Google but actually useful.
- Gemini Pro's Deep Research mode is where things get serious. It goes broad and deep on a topic and — this is the killer feature — you can export directly to Google Docs. That means you get a fully formatted document you can share with your team, turn into specs, or use as a starting point for planning. Super underrated.
3. Learning & Knowledge: NotebookLM + Gemini Gems
This category is a bit different — it's less about writing code and more about staying sharp as an engineer. But honestly, continuous learning is part of the job, and these two tools make it significantly more efficient.
🎙️ NotebookLM
Feed it anything — a YouTube video, a PDF, a website URL, copied text — and NotebookLM turns it into something you can actually learn from. We're talking AI-generated podcasts, presentation slides, mind maps, memory graphs, flashcards, and quizzes, all based on your source material.
Trying to get up to speed on a new framework? Drop in the docs. Want to study for a certification? Throw in the learning materials. It transforms passive content into interactive learning in a way that genuinely sticks. One of the most underrated tools on this list.
💎 Gemini Gems
Think of Gems as your personal learning tutor, customized to how you work. You can create specialized Gems focused on specific domains — system design, algorithms, a new language you're picking up — and it'll guide you through structured learning sessions. It also supports planned daily tasks, so you can build actual learning routines into your schedule instead of just "I'll learn that someday."
If you're a learning-oriented engineer (and you should be), this combo is hard to beat.
4. Specs, User Stories & Planning: Spec Kit from GitHub
Planning is where a lot of teams waste time. Spec Kit helps you generate specs, user stories, and task breakdowns faster — and the best part? It's compatible with basically any AI tool you're already using: Cursor, Claude, and others. No lock-in, no friction.
If you're tired of writing boilerplate Jira tickets and spec documents from scratch, this one's for you.
5. Code Review: CodeRabbit
CodeRabbit does automated AI code reviews, and it's genuinely good. It catches things that slip through human reviews — edge cases, logic issues, potential bugs — and it integrates directly into your PR workflow.
Fair warning: it's a bit on the pricier side. But if your team is serious about code quality, it's worth evaluating.
6. Claude & Claude Code — The Best One Out There (By Far)
Okay, I'll be honest — this is the one I'm most excited to talk about.
Claude (and especially Claude Code) is, in my opinion, the strongest AI tool available right now for software engineers. It's exceptional at deep research, planning, and end-to-end task execution. It doesn't just answer questions — it thinks through problems with you.
But what makes Claude really powerful is the ecosystem you can build around it. Here's what I mean:
MCP Servers, Skills & Plugins
Claude supports MCP (Model Context Protocol) servers, skills, and recently added plugins — all of which let you extend what Claude can do. Some of my favorites:
🎨 Figma MCP Server / Skills
Connect Claude to your Figma files and now your AI actually knows what the design looks like. Instead of generating random UI from thin air, it writes code based on real designs. And it works the other way too — you can generate Figma design files from prompts. Huge for frontend work.
🖥️ Frontend Design Plugin
This one generates production-grade frontend code with real design polish. If you've ever gotten AI-generated UI that looks generically bland, this is the fix. The output is actually distinctive and deployable.
📡 Postman MCP Server
Connects Postman to Claude so you can interact with your APIs using natural language. Manage Postman resources, automate API workflows, and let AI agents work with your actual API layer. If you do a lot of API work, this is a game changer.
⚡ Superpowers
This one might be my personal favorite. Superpowers extends Claude with capabilities like brainstorming frameworks, subagent development, code review, debugging, TDD workflows, and skill authoring. It makes Claude significantly smarter about the process of software development — not just writing code, but doing it well.
One heads up: it's token-heavy, so be mindful of that if you're watching usage.
Conclusion
Claude Code is doing something different from the rest. It's building out a full AI engineering ecosystem — security tooling, code reviews, richer in-chat experiences, and more. It's not just a chat interface anymore; it's becoming the operating layer for AI-assisted software development.
My strong recommendation: invest real time into learning how to get the most out of it. Set up the MCP servers, experiment with skills, and build your own workflow around it. The engineers who figure this out early are going to have a serious edge.
The tools are here. Time to use them.
One more thing — if you want a structured way to get started, Anthropic put together a completely free tutorial platform covering a wide range of topics around Claude and AI development. It's genuinely well done and worth bookmarking: anthropic.skilljar.com
Have a tool I missed? Drop it in the comments — always looking for what's actually working for other engineers.
Top comments (0)