Speed, Precision, and "Memory": The Power User Update

Speed, Precision, and "Memory": The Power User Update

We are moving fast. This week’s updates are all about removing friction and giving you professional-grade control over your workflow. We looked at the bottlenecks where you were waiting or where you needed more precision, and we engineered solutions to blast through them.

Here is the story behind the major improvements now live on Markolé.

The Multi-Tool Visual Assistant: Perfect Alignment in One Move

The hardest part of building a brand is keeping everything consistent. Usually, if you change your color palette, you have to remember to update your logo briefs, your sample images, and your brand data separately. We’ve solved this by upgrading the Visual Assistant to run multiple tools simultaneously.

Now, when you ask the Assistant to "make the colors warmer," it doesn't just change a text description. In a single, fluid motion, it updates the underlying data structure, regenerates the actual color palette swatches, and generates a new sample image to show you how those colors look in real life. It ensures that every element of your visual identity stays perfectly aligned, automatically.

Surgical Control Over Strategy

In the past, regenerating your strategy could feel a bit like using a sledgehammer when you needed a scalpel. To give you more control, we have broken down the Aligned Identity Framework in the backend.

You no longer have to regenerate an entire section if only one part feels off. You can now dive in and regenerate or edit a single specific narrative—like just your "Brand Archetype" or "Mission"—without disturbing the rest of your approved strategy. For those who prefer speed, we also added the ability to trigger multiple synthesis steps at once, allowing you to sprint through the initial creation phase before slowing down for that granular refinement.

A Smarter Brain with "State Memory"

We have migrated the vast majority of our intelligence layer to Google’s Gemini models, but this is more than just a model swap. We are utilizing a powerful feature called "Thought Signatures."

This isn't about showing you the AI's internal monologue; it is about state preservation. The AI now has a much better memory of the previous state of your work. This is a game-changer for multi-turn editing, especially for images. If you ask to edit an image, the AI perfectly recalls the context and details of the previous version, allowing for much more accurate and nuanced adjustments. It feels less like rolling the dice and more like iterating with a human designer.

(Note: While Gemini runs the show for strategy and layouts, we have kept OpenAI in the loop specifically for Logo Generation because it is still the best at handling transparency.)

Faster Generation and Non-Stopping Flows

Waiting for images to generate one by one is a creativity killer, so we’ve rebuilt the generation pipeline. When you create Presentations or Marketing Documents, images now generate in parallel with a smart stagger. We also generate placeholders instantly, so if a specific image fails, the layout doesn't break—you can simply regenerate that single slot.

Finally, we’ve smoothed out the error handling. If a minor issue occurs, it no longer interrupts your flow. Errors are logged into a persistent notification center, allowing you to check them out later without losing your creative momentum.

These updates are live now. Log in, try the new flow, and let us know how it feels.