Tools I'm Using Right Now - Feb 2026 Edition
These aren’t the “best tools of 2025-26.” This is just what’s in my rotation right now. As an AI engineer, the toolkit shifts fast — what I swear by today might get replaced next month (or next week). That’s part of the fun. So consider this a snapshot, not a recommendation list.
Claude Code
If you haven’t tried Claude Code yet, go do that. I’ll wait.
The CLI-based agentic coding experience from Anthropic is something else. OpusPlan is the standout — the 4.6 models are absolutely killing it. The ability to create sub-agents directly in the CLI is a great feature, and agent skills slot in naturally to how I already think about breaking down problems.
The one real downside? It’s so good that hitting token limits becomes a genuine part of the planning process. You start thinking about token budgets the same way you think about compute budgets. That’s a weird problem to have, but here we are.
Gemini Antigravity
The best agentic IDE, in my opinion. Full stop.
The ability to switch between Gemini and Claude models quickly is amazing — you’re not locked into one provider’s strengths and weaknesses. Customization, skills, and agent management are all well built into the system. It feels like the tool was designed by people who actually use agentic workflows daily, not by people who watched a demo once.
The recent Gemini 3.1 Pro model made it vastly better. If you tried it six months ago and bounced off, give it another shot.
Astro
This one is a newer discovery for me, and it’s what this site is built with. The minimal client-side overhead makes these sites incredibly snappy — we’re talking static HTML with zero JavaScript bundles shipping to the browser.
But the real win is the developer experience. The syntax and data model are incredibly easy to use. Content collections with the glob loader mean I write markdown files, define a schema, and everything is typed end-to-end. No CMS app needed. I prefer writing directly in markdown anyways, so this removes an entire layer of complexity I never wanted.
If you’re building a content-focused site and you’re tired of fighting your framework, just try Astro. You’ll be annoyed you didn’t switch sooner.
Opal
Another Google product (sensing a theme?), and one that surprised me. The no-code simplicity of the workflows and the ease of using the built-in tools makes this a fun and genuinely useful tool. Now that it’s directly integrated into the Gemini app, it works great for automating the kinds of repetitive tasks that slowly eat your day.
My one gripe: the ability to schedule workflows would be a nice addition, similar to what’s available in Gemini Enterprise. But alas, still a great tool even without it. I’m betting that feature shows up eventually.
Wand
Okay, this one is less of a “tool” and more of a quality-of-life improvement — but it’s earned its spot on this list.
I love gaming. Always have. But as a dad, I don’t have as much time as I used to. And frankly, things like inventory management in games like Skyrim and Fallout are tedious. I said it. Tedious.
Enter Wand, where I can flip on “cheats” to make single-player games more approachable with limited time. Skip the grind, get to the good stuff. Now before anyone gets upset — there’s absolutely no room for cheating in online games. But for single-player RPGs? This has been a welcome addition to keeping gaming alive in the dad era.
WSL
I love developing on a Mac. It’s simple, easy to access the right configurations, easier to manage PATH, lots of reasons. But I hadn’t seriously tried WSL on Windows until recently.
Is it a Mac replacement? Absolutely not. But it’s a surprisingly great dev experience, especially when you’re used to working in VS Code. The WSL integration with VS Code is seamless — it’s the closest thing you’ll get to a Linux experience on Windows with minimal effort. If you’re on a Windows machine and haven’t set this up yet, it takes about 10 minutes and it’s worth every one of them.
Wrapping Up
Tools change fast. Especially in the AI space, what works now might not be what I’m reaching for in six months. But that’s kind of the point — staying curious, trying new things, and keeping what actually makes you faster. I’ll probably do another one of these when the stack inevitably shifts.