Over the past year, I’ve grown increasingly dependent on a quiet set of AI tools—most of them free, all of them underrated. While everyone was rushing to try the same mainstream models, I started building a creative workflow that feels more like a collaboration between me and a network of specialized assistants.
AI in 2025 isn’t just about automating tasks or generating images—it’s a mindset shift. The right tools don’t replace your creativity, they provoke it. They challenge your assumptions, expand your vocabulary, and offer unexpected angles on what’s possible.
Here are seven tools that fundamentally shifted the way I approach my work as a designer.
1. ComfyUI

I’ll be honest: ComfyUI is not for casual users. It’s a node-based interface for Stable Diffusion that lets you build your own custom generation pipelines. But once you get over the initial learning curve, it becomes incredibly powerful. You can design complex flows that control every aspect of image creation—from the model used to the order of processing, from text prompts to face correction.
I started using ComfyUI to develop visual identities for experimental projects—like exploring multiple aesthetic directions with fine-tuned parameters. Over time, I realized I could reuse my own workflows like templates, adjusting just a few values for entirely different outcomes.
Why I use it: Because it gives me complete control over the generative process, and I can build repeatable visual pipelines that evolve with my style.
2. Recraft

Recraft was made for designers who want more than pretty pictures. It’s the only free tool I’ve found that consistently generates clean, editable SVGs with brand-ready typography and layout control. Unlike most text-to-image tools, it understands design structure—grids, alignment, white space.
I often use Recraft to quickly create posters, social graphics, or presentation visuals. The typography is surprisingly strong for an AI tool, and it’s one of the few that I can trust to generate text that doesn’t need post-editing. It also offers an “AI Suggest” function that I use to explore color palettes or stylistic variations.
Why it matters: Because I need assets that are not just creative, but also usable, editable, and production-ready.
3. Ideogram

When I first tried Ideogram, I didn’t expect much. But it solved a problem that’s plagued every other image generator I’ve used: legible, aesthetically integrated text. Ideogram is great when you want visuals with embedded type—think infographics, mock app UIs, or educational content. The model is designed with an understanding of how words fit within a visual composition.
I’ve used it to create quick diagrammatic illustrations for presentations, to mock up labels, and even to experiment with graphic t-shirt concepts. The output feels intentional, not AI-glitchy.
Why I recommend it: Because it’s the first AI image tool where the text actually works—visually and semantically.
4. Flux

Flux is less known than tools like Runway, but it’s incredibly powerful when you want to iterate inside a visual composition. Unlike tools that make you start from scratch every time you edit, Flux lets you isolate and prompt-adjust parts of the image. You can revise just the lighting, or just the background, or just the character expression—without losing the rest of the composition.
This kind of precision feels more like art direction than image generation. I’ve used Flux in branding projects where consistency across frames matters, and in speculative storytelling where I want to evolve a scene without destabilizing it.
Why I keep using it: Because it supports creative iteration—not just generation.
5. Kittl

Kittl is like if Canva had a baby with Illustrator—but smarter. It’s web-based and AI-supported, but built with designers in mind. The standout feature for me is its typography engine. Kittl generates layered, expressive text compositions that look like they were handcrafted by a professional letterer.
When I’m brainstorming packaging layouts, poster series, or merch concepts, I go to Kittl first. It lets me explore layouts rapidly and export vector assets without friction. It’s also helpful for client presentations when I need polished visuals fast.
Why it earned a spot in my toolkit: Because it helps me sketch in high fidelity—especially when typography is the hero.
6. PromptCanvas

Unlike most of the tools on this list, PromptCanvas doesn’t generate images. It generates structure. It’s a writing and ideation space where you break down your prompt into logical components—goal, tone, context, format—and then remix or expand each piece.
I use it to write copy for landing pages, product ideas, and design briefs. It’s a way of thinking that slows you down just enough to be intentional, while still letting AI offer quick suggestions and alternatives.
Why I use it before I design anything: Because it helps me clarify the “why” before I dive into the “how.”
7. POET

POET is a more experimental tool—but probably the most creative on this list. It’s a prompt expansion engine that shows you the unexplored edges of your idea. Instead of refining one version, it explodes your concept in different directions. The results are often weird, surprising, and sometimes exactly what I didn’t know I was looking for.
When I’m stuck or bored with my own design ideas, I’ll throw a prompt into POET and just read through what it suggests. A phrase, a mood, a contradiction—it only takes one unexpected shift to unlock a new path.
Why I return to it when I feel uninspired: Because POET doesn’t just answer—it provokes.
Final thoughts
Every one of these tools serves a different purpose in my creative life. Some are technical, some are expressive, some are introspective. But all of them have one thing in common: they push me to design more intentionally.
We’re living in a time when AI can either flatten our originality or amplify it. The difference lies in how we use it. So don’t just follow the hype. Experiment, combine, iterate. Let these tools challenge you.
And if you end up discovering a new one that belongs on this list, send it my way. I’m always looking for new collaborators. Even the machine kind.