My AI tool stack for 2026 🛠️

Hello everybody, welcome back to Moodboard 🌴

Last year, I tested a lot of AI tools.

Some I tried once and never touched again. Others I played around with for a couple weeks before losing steam. 

There were a handful, however, that really stuck and became foundational to my daily work.

In this week’s edition, we’re walking through the 5 AI tools that I'm using every single day (along with some workflows, prompts, and resources to help you get started with each).

Let's dive in 👇

1. Flora

This first one is a tool I haven't mentioned in Moodboard yet.

I only started using it recently (shoutout to Mike Wigmore who shared it during one of our Vibe Marketer OS office hours last month), but it has quickly become the home for many of the multi-modal workflows we use across both our agency and media company.

Flora is an ‘node-based’ AI workspace that lets you chain together different AI models, tools, and data sources into repeatable workflows.

Think of it like a visual command center where you can build automations that pull together actions from different LLM’s, image generation models, upscalers, and more, all in one place.

This has been huge for me and my team.

Instead of jumping between tabs and copy-pasting outputs from Claude to Nano Banana Pro to Leonardo and so on, Flora lets you build the entire workflow once and run it on repeat. 

We’ve been porting many of the prompts and processes I've shared in Moodboard over to Flora, and it has been a very natural transition.

Keep an eye out for a more detailed Moodboard guide to Flora very soon!

2. Midjourney

Still the most aesthetic, personalizable image generation option out there.

I've been using Midjourney since 2022 (back when you could only access the tool via Discord), and it has continued to evolve to stay on the cutting edge of realism and aesthetics.

With Midjourney's V8 model rumored to be right around the corner, this tool is (somehow) only going to get better.

Where Midjourney shines:

• Personalization: Midjourney learns your aesthetic preferences over time and gives you a personalization code that biases outputs toward your taste in a very nuanced way. This alone makes it worth using.

• Moodboards and style references: You can feed Midjourney reference images to dial in a specific look. Upload a moodboard of visuals you like, and it will pull from that combined aesthetic when generating new images.

• Aesthetic superiority: When the goal is "make this look good" rather than "render this text accurately," Midjourney is still unmatched.

I've published a lot about Midjourney in Moodboard, but if you're looking to get started, check out our "How to Get the Most out of Midjourney" piece.

3. Claude

This one should come as no surprise. Claude has been my LLM of choice for all of 2025.

With the addition of Claude Skills, persistent memory across chats and projects (a hugely underrated update), and Opus 4.5 (the most impressive model for my money across any of the frontier labs), Claude is still my daily driver.

Claude's writing quality is noticeably better and more human than other models. The outputs require less editing, and the Projects feature means I can set up dedicated workspaces for different clients, use cases, and deliverables.

If you want to see how I use Claude Projects to write newsletters faster, I walk through my entire setup in this YouTube video.

4. Nano Banana Pro

I'll be honest. Up until recently, I didn't use Gemini at all.

Claude has always been my LLM of choice, and for image generation I'd end up using ChatGPT or Midjourney depending on the use case.

Nano Banana Pro has completely shifted the paradigm.

Whereas I use Midjourney for branding and more aesthetically-minded image generation, Nano Banana Pro is currently the best solution for creating the text-heavy imagery that marketers/creators/founders/etc. spend so much time putting together. Think infographics, social posts, explainers, quote graphics, and anything where you need readable text baked into the image.

The text rendering is dramatically better than any other image generator I've tested. I'm now creating assets in minutes that used to require Canva and 45+ minutes of manual work.

If you want to try it out for yourself, I wrote a full breakdown of our ad creation workflow within Nano Banana Pro that you can check out here.

5. Wispr Flow

Voice-to-text is truly the optimal medium for interacting with AI.

It removes all of the friction of typing long, perfectionist prompts and allows you to ramble freely with confidence that the LLM will understand what you're looking for.

Wispr Flow is my favorite solution for this, and I wear my power user status as a badge of honor.

When you're typing, you self-edit, overthink word choice, and generally just try to sound smart.

When you're speaking, you just... explain what you need. And that natural explanation is often all that AI needs to give you a good response.

If you're still typing during all of your conversations with LLMs, you're leaving a lot of speed, iteration, and quality on the table.