It's Monday. Here's what changed in AI this week, and what it means for your design work:
🧪 What I'm Building: Limora V2 is live
🚨 Big News: Limora, the first AI asset generator built for designers
🤖 Prompts Inspiration: Hero Images and Animations With the Exact Prompts
⚙️ Tool of the Week: World Labs Marble 1.1 Plus
🔗 Quick Links: 5 things that caught my eye
WHAT I’M BUILDING
Behind the Scenes
Last week I launched Limora V2.
I'd quietly run the MVP for four months. Just enough to validate the idea, talk to designers, and figure out what V2 needed to be.
Then we shipped. The first two days were stressful in a good way - the launch video went viral on X way harder than I expected. We had to scramble to handle the traffic, the support, and the feedback all at once.
The response has been incredible. First paid users are in. Detailed feedback rolling in. A clear list of features we're already building next.
Hitting that first paid user milestone after four months of building is a feeling I'll remember for a long time.
Massive thank you to Danilo, who carried the entire product development side, and Stephen, who led the design. Limora wouldn't exist without them.
This is the start. Doubling down from here.
BIG NEWS
Limora V2, the first AI asset generator built for designers
Every other AI image tool was built for prompt engineers. Limora was built for designers - and that's the entire point.
Limora V2 went live last week. It's the first AI asset platform built around designer workflows, not generic prompt boxes. Instead of one blank input, you get 14 dedicated workflows: hero images, illustrations, icons, product photos, desktop mockups, logo variations, image animation, and more.
Each workflow is tuned for one asset type. The controls, references, and presets match what designers actually need - not what a prompt engineer would type.
The other big shift in V2: it's no longer just for web designers. Brand designers, product designers, graphic designers - if you ship visual assets for clients or products, Limora covers your day-to-day.
My take
Every other AI image tool drops you into a blank chat and expects you to know how to prompt. That's exactly why most AI-generated assets still look generic. Limora flips it. The workflow does the heavy lifting. You bring the direction, the style, the taste. The output feels designed, not generated.
The bigger play: this is what kills the AI slop in professional design work. When the tool is opinionated about how an asset should be built, the output stops looking like everyone else's.
How to use it
Start with the asset type you ship the most - probably hero images or icons. Lock in a reference style. Generate three options. You'll see the difference in the first round.
PROMPTS INSPIRATION
Hero Images and Animations With the Exact Prompts
Background Image
Surreal 3D-rendered dreamscape of rolling, velvety moss-covered hills under a soft, hazy sky. A winding, illuminated stone pathway made of smooth flat slabs cuts through...3D Head Avatars
A hyper-stylized, dynamic 3D head portrait floats against a vivid red background with soft volumetric clouds and seamless shading. Only the cartoonish...Hand Drawn art
Hand-drawn marker illustration of [subject], bold black outlines, saturated vibrant colors, visible marker strokes, playful and energetic composition,...TOOL OF THE WEEK
World Labs Marble 1.1 Plus
You can already generate images. You can generate video. But generating an entire 3D world you can fly through - that's new..
Marble 1.1 Plus is World Labs' latest update to their multimodal world model. Feed it a text prompt, a single image, multiple images, a panorama, or a coarse 3D layout - and it generates a fully explorable 3D scene. Export as Gaussian splats, polygonal meshes, or rendered video.
1.1 Plus is the version built for scale - bigger, more complex worlds, automatic expansion across up to 5 dynamic cubes in a single generation.
🧠 How I'd use it
Concept environments for brand work. Generate the world your product lives in - then drop the product in via 3D render or video composite.
Background plates for hero videos. Skip 3D modeling entirely. Generate the scene, export as video, layer your UI on top.
TUTORIAL OF THE WEEK
Fully Automate Email Design With Claude Code + Resend
Email design used to mean hand-coding HTML or fighting drag-and-drop builders. This workflow ships an entire branded email sequence in one sitting.
Step 1: Generate the copy in Claude
Open Claude and describe your email sequence
Let it write the full copy
Save the output to a markdown file
Step 2: Build the full email sequence
Open Claude Design, upload your brand's design system, and generate React code for the first email
Drop that template and the markdown into Claude Code and ask it to generate the rest of the emails from the same design
Iterate until every email looks right - use Limora for custom graphics that make them stand out
Step 3: Ship with the Resend MCP
Use the Resend MCP inside Claude Code
It pushes all your React emails to Resend and sets up the automations for you
The full sequence is live without leaving the terminal
QUICK LINKS - This Week's Radar
Figma Make goes to beta with direct codebase editing and PR workflow - Select any UI element on screen and Make finds the code, edits it, commits, and opens a PR.
Higgsfield drops "Hell Grind" - the first fully AI-generated 95-minute feature film, built in 2 weeks on Seedance 2.0.
AI in Design 2026 report - 900+ designers, 60+ countries. Half have shipped AI-generated code to production.
Midjourney V8.1- HD by default, native 2K rendering, 5x faster than V8.
Anthropic opens Milan and Seoul offices- first major EU and APAC enterprise push.
How was this week's newsletter edition?
💌 Know a designer who should be using AI smarter?
Forward them this email. Or just send them to logiaweb.net to join.
See you next Monday,
