Product

The latest updates from the Playground team

Up until now on Playground, most templates and generated images were basically flat pictures. You could ask the AI to redo the whole thing with a new prompt, but you couldn't just reach in and change one specific thing, like the headline, or the logo, or that one little element that wasn't doing it for you.

Smart Layers changes that. Every image in Playground is now broken into individual, editable layers that you can select, move, resize, rotate, delete, or rewrite. Stack a few layer edits together, type a prompt, and Playground regenerates your design with all of your changes baked in.

Here's how it works, why we built it, and what it looks like in practice.

## Why we built Smart Layers

For a long time, the only way to change an image was to prompt the AI to do it. "Make the headline say Bachelorette." "Move the grapefruit a little to the left." Sometimes it worked. 

A lot of the time it'd change the thing you asked about but mess with something else too. The font would shift, the layout would move, you'd end up further from what you wanted than when you started.

We wanted two things at once: the ability to make precise, surgical edits to individual pieces of a template, and the ability to let AI take those edits and run with them. Smart Layers does both.

## A real example: Paloma Party to Bachelorette Beach

<video 
  src="https://cdn.playgroundai.com/whats-new/segments/smart-layers.mp4" 
  width="100%" 
  style="border-radius:24px;"
/>

The best way to explain Smart Layers is to walk through an actual edit. Say you find a "Paloma Party" cocktail template that has the right energy for what you're making, but you need it to be a bachelorette party invite with a tropical feel.

Here's how that goes.

**Select a layer.** You click the grapefruit garnish on the cocktail glass. The rest of the design dims, and the garnish highlights as its own layer with a small toolbar above it: edit, duplicate, delete. Every visual element in the template works this way. The text, the glass, the decorative sparkles, the background. Each one is its own selectable, editable piece.

**Move and rotate it.** You grab the grapefruit and drag it to a new position. Pull on the corner handles to resize. Twist the rotation grip to tilt it. The rest of the design stays exactly where it is. This edit gets saved as a step, shown as a small chip at the bottom of the screen so you can see what you've changed so far.

**Edit the text.** You click the "PALOMA" text layer. It opens up as an editable text field right on the canvas. You clear it and type "Bachelorette" instead. Hit confirm. That edit gets added as another chip next to your first one. The design now shows both changes: the repositioned garnish and the new headline text.

**Prompt AI with your edits as context.** Now here's the part that makes Smart Layers more than a basic layer editor. At the bottom of the screen, you can see your stacked edits, and there's a prompt field. You type: "Tropical themed with palm trees, beach." Hit send. Playground takes your layer edits and your prompt together, and regenerates the design.

**The result.** The "Paloma Party" template has become a "Bachelorette Party" invite with a full tropical beach scene: palm tree silhouettes, a sunset sky, sandy shore, the cocktail glass still front and center with the retro typography style preserved. The text reads "Bachelorette Party." The layout keeps the same bones, the same feel, but the context is completely different.

That whole sequence took maybe 30 seconds.

## What makes this different

Plenty of tools let you move stuff around on a canvas. Smart Layers is doing something more specific. Each edit you make to a layer becomes an instruction that AI can work with.

Move the garnish, change the headline, type a prompt about a new theme, and the AI doesn't just blindly regenerate from scratch. It takes your specific layer edits as input. It knows you moved the garnish. It knows you changed "Paloma" to "Bachelorette." It reads your prompt about palm trees and a beach. Then it produces something that respects all of those decisions at once.

This is the difference between "regenerate and hope" and "direct the AI with your hands and your words at the same time."

## Everything you can do with a layer


### Select

Click any element in a template. It highlights with a selection box and handles. A small toolbar appears above it with options to edit, duplicate, or delete. The rest of the canvas dims so you can see exactly what you've grabbed.

![demo gif](https://cdn.playgroundai.com/smart-layers/select-an-object.gif)


### Move

Drag any selected layer anywhere on the canvas. Guides appear to help you stay aligned with the rest of the design.

### Duplicate

Need another copy of an element? Select it and hit duplicate from the toolbar. Useful when a design needs repeated motifs or when you want to try a variation without losing the original.

### Resize

Pull corner handles to scale proportionally. Pull side handles to stretch in one direction.

### Rotate

Grab the rotation handle above the selection and twist. Hold shift to snap to clean angles like 45 or 90 degrees.



### Delete

Don't need an element? Select it and delete it from the toolbar. The rest of the design stays put. Maybe the template has a decorative graphic that clashes with your brand, or a subtitle you don't want. Just take it out.

![demo gif](https://cdn.playgroundai.com/smart-layers/resize-edit-rotate-delete-duplicate-objects.gif)

### Edit text

Double click a text layer and it opens as an editable field right on the canvas. Type your new words. The font, styling, and layout of the original carry over after you send in the edit.
![text-actions-on-elements.gif](https://cdn.playgroundai.com/smart-layers/text-actions-on-elements.gif)

## Stacking edits and prompting

![edits-can-be-stacked](https://cdn.playgroundai.com/smart-layers/edits-can-be-stacked.gif)


This is where Smart Layers really separates itself from a standard layer editor. Every edit you make, whether it's moving a layer, changing text, or deleting an element, shows up as a chip at the bottom of the screen.

So you can make a few specific changes by hand, then describe what you want the whole design to feel like, and the AI puts everything together. Your edits give it the specifics. Your prompt gives it the bigger picture. What comes out reflects both.

You don't have to add a prompt at all if you don't want to. If you're just swapping a name on an invitation, do the swap and you're done. The prompt is there for when you want to take the design somewhere new.
## When Smart Layers is the right tool

Smart Layers is built for the work that used to make you start over for no good reason. Swapping a headline or event name. Dropping in your logo or product photo. Repositioning elements that are almost right but not quite. Removing pieces you don't need. Then optionally telling AI to take your changes and shift the whole mood, theme, or setting.

## When to start fresh instead

If you don't like anything about the template, not the layout, not the style, not the composition, Smart Layers isn't going to get you there. It's designed for designs where the structure is right and you need to adjust the details, then optionally push the whole thing in a new direction with a prompt. If the bones themselves need to be different, pick a new template or start from a blank canvas.

## Tips

**Edit the layers first, prompt second.** Get your specific changes in place (text, position, deletions) before you write a prompt. This gives the AI more to work with and produces better results than a prompt alone.

**Watch your edit chips.** The chips at the bottom of the screen are your running list of changes. If something isn't coming out right after a prompt, check what edits are stacked. You might have an old move or deletion in there that's sending things in a direction you didn't intend.

**Use templates as starting points.** The whole reason Smart Layers exists is that templates are meant to be jumping-off points. Pick one with the right energy, then shape it into exactly what you need. Hunting for a template that already matches every detail is slower than finding one that's close and editing it.

## Try it

Open any template in [Playground](https://playground.com/design/explore), click on a layer, and start editing. Move something, change the text, delete what you don't need. Then type a prompt and see what comes back. The whole system is built to reward experimenting, and nothing you do is permanent until you want it to be.

Smart Layers: Unlocking interactivity in every image

Is Playground AI Free? Here's Exactly What You Get (2026)

Typography plays a huge role in how a design feels. A small change in font can completely change the tone of a poster, a social graphic, or a piece of merch.

We’ve expanded Playground’s font library with **~1800 fonts**, giving you **over 20× more options than before**. Whether you're designing something clean and minimal or bold and expressive, you now have a much wider range of type to work with.

Along with the expanded library, we’ve added a few tools to make finding the right font much easier.

---

## Search Fonts by Vibe

![vibe](https://images.playgroundai.com/user-uploads/cmhxhgcm702guumc2okoul1y2/cmmo1sr5j016w01s6c7mn5f5k.png)

Sometimes you don’t know the exact font name you’re looking for. You just know the feeling you want the design to have.

Now you can **search fonts by vibe, style, feeling, theme, or whatever you are envisioning!**

Try searches like:

- childrens themed
- retro diner
- vintage western
- elegant wedding
- fun summer party
- elegant luxury fashion brand
- cozy christmas holiday card
- retro 70s disco poster

This makes it much faster to explore fonts that match the style you're going for without scrolling through hundreds of options.

---

## Find Similar Fonts Instantly

![similar](https://images.playgroundai.com/user-uploads/cmhxhgcm702guumc2okoul1y2/cmmo1su3q016z01s67b3kfc1h.png)

When you find a font you like, you often want something close to it, but slightly different.

You can now **view fonts similar to the one you're currently using**. This makes it easy to compare options, try subtle variations, and quickly find the version that fits your design best.

---

## See Suggested Fonts for Your Design

![suggested](https://images.playgroundai.com/user-uploads/cmhxhgcm702guumc2okoul1y2/cmmo1st3d015801s61fm30khy.png)

Fonts don’t exist in isolation. The right font depends on the layout, the graphics, and the overall tone of the design.

Playground now **suggests fonts based on the design you're working on**, helping surface options that pair well with your layout and style.

It’s a simple way to discover fonts that fit naturally with what you’re creating.

---

## Try the New Fonts

With thousands of fonts and better discovery tools, it’s now much easier to experiment with typography and find the right style for your design.

Browse templates and try the new fonts in your next project:

https://playground.com/design/explore

Fonts Just Got a Big Upgrade

We’ve just rolled out **Extract Text**, a new tool in Playground that lets you turn text from any template into editable layers.

Now, even templates that were previously flattened can be fully customized. 

Change the text, adjust fonts, modify spacing, and style it to fit your design, all without recreating anything from scratch.

---
## How to Use Extract Text

Using the tool is straightforward:

1. **Select a template** – Any template on Playground works, regardless of whether it already has layers.
2. **Run Extract Text** – The tool converts the text into editable layers.
3. **Edit your text** – Adjust the wording, font, color, size, or spacing directly in the editor.

**Note:** Running Extract Text will use credits from your account. The process usually takes 30–60 seconds while the system analyzes the image and rebuilds the text as editable layers.

![Frame 2147221308](https://images.playgroundai.com/user-uploads/cmhxhgcm702guumc2okoul1y2/cmmjn4il0012001s63jbr0s9j.png)

----

## When the Extract Text Tool Works Best

The tool performs best with **simple, clean text**.

![Frame 2147221305](https://images.playgroundai.com/user-uploads/cmhxhgcm702guumc2okoul1y2/cmmjltnzt00dv01s63nezezru.png)

Things like:

- Straightforward typography
- Single-color text
- Basic layouts without heavy effects

In these cases, the system can usually detect the text and recreate it as an editable layer fairly accurately.

Where it struggles is with **more complex typography**, such as:

- Curved text
- Heavy shadows or outlines
- Multi-color text effects
- Decorative borders or gradients

![Frame 2147221306](https://images.playgroundai.com/user-uploads/cmhxhgcm702guumc2okoul1y2/cmmjltqym00dy01s6gd8sdwn8.png)

These styles make it harder to accurately reconstruct the original text.

That said, even when the result isn’t perfect, it can still give you a **strong starting point** that you can quickly refine.

---

## Example: Before and After

Below is an example of how Extract Text can be used in practice.

The first image shows a design where the text is completely flattened into the image.

After running Extract Text, the design becomes **properly layered and editable**.

![Frame 2147221307](https://images.playgroundai.com/user-uploads/cmhxhgcm702guumc2okoul1y2/cmmjn1wsv010q01s63kkfb7iq.png)

---


Editable text layers make templates far more flexible. You can update text without starting over, adjust designs quickly, and keep everything looking clean and consistent.

👉 Try it for yourself, pick any template and see how easy it is to edit: [Explore Templates](https://playground.com/design/explore)

Introducing Extract Text: Turn Flat Text Into Editable Layers

It’s with a warm heart that we’re excited to announce a new team is shaping the future of Playground. No the rumors aren’t true, we aren’t going away or slowing down in the slightest. The Playground you know and love is here to stay and we are just getting started!

So who are we? We’re a team of entrepreneurs with diverse startup backgrounds motivated by the future of AI and how it will redefine the human experience as we know it. Over the last month since the acquisition was finalized, we’ve grown close to the community and wake up every day to serve you to the best of our abilities. We are devoted to putting users first and delivering an experience that is magical.

We’re standing on the shoulders of giants and are immensely grateful for the previous teams efforts. They are innovators at heart who helped us foot in step during the transition process to minimize disruptions. Special thanks to Patrick, Justin, Alex, David, and Suhail — you guys rock!

There is more to come, we have some exciting product updates we’re looking forward to sharing with you soon. For now, we want this letter to serve as something steady and clear: a commitment that we will prioritize our community, communicate openly, and continuously put our best foot forward. We welcome any and all feedback from our users: candid, unfiltered, frustration, praise, we want to hear it all. If you’re interested in sharing your thoughts with us, please [reach out here](https://tally.so/r/5BXxpE).

If you care deeply about creative tools, thoughtful design, and building in the future of AI, we’d love to meet you and see if you’re a good fit for our team. You can learn more at [our hiring page](https://www.notion.so/Work-with-us-2a98b80b88dd80a5b068e851c2919b47?pvs=21).

Thank you for your patience, passion, and continued presence. We’re grateful to be part of the next chapter with you.

Onwards and upwards,

The Playground Team

Playground's Next Chapter

Our newest AI model Playground v3 (or "PGv3") is purpose-built for graphic design capabilities. 
We focused on being the best in the world at prompt understanding and control – pushing beyond aesthetics, which has become saturated as a benchmark. PGv3 outperforms all the most popular image models made in its class, as show in our [technical report](https://arxiv.org/html/2409.10695v1).  

Creating an AI graphic designer also requires introducing new model capabilities. Our research focused on improving control over visual communication, typography, color precision, and layout. We also introduce Argus, a new vision language model that can describe images better than a human could (or would want to). VLMs are a key component to improving our image model so that people to control every element of a design.
To learn more about PGv3's capabilities and performance, go to our [PGv3 landing page](https://www.playground.com/pg-v3). And for more specifics into our research methods, evaluations, and benchmark testing, please see our technical report on Arxiv. 

We're also hiring: [playground.com/jobs](https://playground.com/jobs)