Napkin AI wants you to stop using so many words in your PowerPoints and Google Slides.
Napkin is a text-to-visual AI tool where you can upload text to a blank doc, or a "napkin," and it will generate a visual representation of your information. Napkin is one of a plethora of AI startups that sprung up as generative AI technology advanced with chatbots like ChatGPT and image generators like Midjourney. It's not a data visualization tool -- you can't upload an Excel database and get bar graphs and pie charts. But you can use Napkin to create graphics for presentations, blog posts, reports and school assignments, among other things.
Napkin officially launched in August, but my colleague and I snagged a preview demo with its founder, Pramod Sharma. Sharma spent a good bit of his career working at Google, specifically on the Docs team, and you can see that influence in the way Napkin looks and functions. Sharma described Napkin as a "visual Google Docs editor," with the goal of users experiencing the "art of making things fun and delightful." Napkin is currently available in English and announced today that it will begin rolling out support for German, French, Japanese and Chinese.
There are a lot of different graphic design and visual creation programs, but giants like Adobe and Canva make up a big part of the market. They're also both investing in AI-powered tools, including image generators like Firefly and Magic Media, with varying degrees of success. I've used and reviewed a lot of AI creative services, and after using Napkin, I do think there's a place for it among these competitors.
The best thing Napkin has going for it is that it's easy to use and can create graphics quickly. It's great for enhancing a presentation quickly or adding extra flair to a report. Students, small businesses and other creators who don't want to spend the time (or money) using advanced services will benefit from Napkin's bare-bones efficiency.
Here are three of the best ways I found to use Napkin. And if you create an account now while it's still in beta, you'll get more personalization options for free.
In my testing of Napkin, I had it create graphics to improve presentation slides, break down a complicated topic for an article and brainstorm different visual ideas. Napkin was good for all the above tasks, but it was by far the best for polishing off a presentation.
A general rule of thumb while creating decks is to have as few words as possible on your slides. This is where Napkin can be helpful. You can upload a chunk of text to Napkin, and it will trim it down and convert it to a visually appealing graphic you can copy and paste onto your slide.
Generation time is quick, between 10-30 seconds. Napkin does collect and anonymize users' data and interactions with its AI for training purposes, but Sharma confirmed via email that your personal information remains private and your uploaded text isn't used to generate content for other users. As with any AI service, I caution against uploading sensitive information to the site.
I pulled an old presentation from a film class I took in college to test it. My original slide was wordy, with seven long bullet points. Napkin came up with really unique graphics, including a pro-con list I really liked and a more artistic visual that summarized my main point of the slide, which was that Netflix's original movies suck. I recreated the slide with the new graphics to compare, and it's way easier to digest now, rather than having people's eyes glaze over with a big block of text.
I pulled some other slides from presentations, including drier ones explaining different theories of communication and summaries of various novels. Each time, Napkin came back with graphics that not only made my slides more fun but also easier to understand. I did do some editing to make my points clear, but for the limited amount of text I took out of the context of the whole presentation, Napkin held up well. One thing I would've loved is the ability to customize my graphics to match the exact colors of my deck theme, but it was all right for now.
If you're not well-versed in graphic design strategy, you might want to take advantage of Napkin as a brainstorming tool. Throughout my testing, I found it to be very creative and versatile, so you can explore a lot of different options based on the same text. To put its creativity to the test, I had Napkin come up with four different graphics to explain the concept of the fediverse, a set of interconnected, decentralized social media platforms built on a set of rules called the ActivityPub protocols. I have a whole article explaining what it is, but I wanted to see what Napkin would come up with to break down a complicated topic. These are some of the top choices.
The first three suggested graphics looked nice, but as someone with an understanding of the material, they're misleading. The first pinwheel graphic correctly identifies several fediverse platforms I list in my text, but incorrectly places the ActivityPub protocols with them; they're what underpin all the platforms. The second graphic pulled a bunch of relevant keywords, but it's not helpful without explaining why they're relevant. The last cycle graphic was interesting because it did pull more implicit information out of my text, but it didn't answer my core question of what is the fediverse.
After scrolling a bit, I found a good match. With this graphic, Napkin correctly pulled relevant platforms, contextualized the ActivityPub protocols correctly and highlighted two key facets of the fediverse -- decentralization and open source. This is a great visualization because it gives the reader a good framework of understanding what's important at a glance and can spark interest to read further. While it took a bit of work to get the best graphic -- and I would've edited it before publishing -- I was pretty satisfied with Napkin's ability to read and synthesize the text I gave it. No AI tool is perfect, after all.
If you're not quite sure what kind of graphic you want, Napkin is great for brainstorming ideas. I pulled some text -- the most I tested was about 157 words -- from a guide I wrote about how to write the best AI image prompt and fed it to Napkin. I wanted to highlight the three essentials each AI prompt should include, but I didn't care how exactly it did that. Napkin came back with a lot of different ideas, and it was nice to be able to browse through all the options since I wasn't married to a specific design or look. This is the one that immediately caught my eye.
Beyond just highlighting the elements I discussed in my text, Napkin's graphic reinforced my point about how essential these things were by displaying them as columns. Without one, the building falls, and the prompt fails. Any of the visuals Napkin created would've been fine, but this one communicates their importance implicitly. This is a great example of something I've found time and time again while testing AI image services -- humans have to be involved in the design strategy and process, otherwise your results won't resonate with your audience. Napkin, like every other AI service, is just a tool to help get you there. For all of these examples, our creators at CNET could've come up with these graphics and probably improved upon them. But for people without a large creative team or a lot of time to create, or for those just looking to spark some initial ideas, Napkin is a great place to start.
It's safe to say that AI art startups like Napkin are in a tough position of having to prove they have something different and useful to offer. In this case, I'm optimistic that Napkin is on track. It's user-friendly and creative enough to provide for amateur creators like me, but it's also a helpful, back-pocket tool for professional creators beginning the design process.
My two biggest wish-list items are more customization options, especially custom colors and fonts, and integrations with other text-process software. I would also love to see Napkin give users the ability to opt out of AI model training entirely, something that I imagine would be desirable for businesses as well as privacy-minded folks.
One of the things that I have found repeatedly with AI creative tools, including Napkin, is how present the human using the service has to be and how these tools can't be used successfully without human-design strategy and decision-making. It doesn't matter if you generate hundreds of different Napkin graphics if you don't know your audience and material well enough to decide which will be the best choice to deliver that information. With that in mind, Napkin is promising, and I'm excited to see it continue to build out new features and functionalities.