What was the catalyst for launching Recraft?
In the summer of 2022, image generation became a thing, and it was obvious that it would transform the whole design industry. This was a rare chance to be a part of a significant change, and to help shape it in the years to come.
My younger sister is a designer, and she guided me through the design workflows, and helped me to learn what designers spend their time on.
I sat down with her and watched her work. One of the things that caught my attention was that she was spending quite a bit of time on tasks like tracing images, which I immediately recognised should be done automatically.
I learned about the importance of vector art and vectorisation for design, and decided to start with solving this need. So, we started by building the first algorithm in the world to generate vector art.
I learned that to use generative AI in design, it is necessary to provide users with a lot of controls over the output, particularly with regards to having the ability to control style, and generate images in their brand style or colours. I also recognised the need to give them style-consistent generation and the ability to generate sets of illustrations, as opposed to single illustrations. And finally, to enable them to iterate on an image to improve and change it according to their needs and vision.
And we’ve been working on innovating in this area since the start of the company.
Tell me about the business – what it is, what it aims to achieve, who you work with, how you reach customers and so on?
Recraft is built for both designers and business users with little-to-no design experience, like marketers or management teams. And it solves the need to create graphic visuals for professional needs, be that illustrating a website or an app, designing social media posts and ads, or creating a brand identity.
And we are continuing to innovate so that the number of tasks Recraft will be able to complete will grow with the increasing quality of image and graphic design generation.
We offer a free and a subscription service, the latter of which has faster image generation and the images remain private, so don’t appear in the community gallery. In the future, we may offer a business subscription, which could be used by a firm.
How has the business evolved since its launch? When was this?
We launched on the 31st of May 2023 on Producthunt. The launch was an immediate success with lots of people using the tool.
We have not done any marketing until now, as we announce our Series A funding announcement. It’s been pure word of mouth. But we still have thousands of users coming to the product weekly. And right now, we have close to 400 thousand registered users.
Tell us about the working culture at Recraft
If I had to describe our culture, there are three main components:
- We ship things. We are shipping every week – often more than once. Updates to the model, updates to UI and UX, stability improvements, etc.
- We talk to each other. I discuss with the team the changes in the product, the strategy, the plans. I encourage feedback, and listen to them. Linked to this, I encourage the team to work from the office so we can collaborate closely together.
- We value the great work of others in the team
How are you funded?
Recraft has completed two rounds of funding. The Series Seed was led by Khosla Ventures. After Seed, we had SAFE-s from several angel investors. And recently, we had a larger Series A round, led by Khosla Ventures and Nat Friednam, ex-CEO of Github, with the participation of RTP Global, Abstract VC, BasisSet, and a number of angel investors.
What has been your biggest challenge so far and how have you overcome this?
I think our biggest challenges lie ahead – and I’m ready for them! The most important thing is not to give up even if it seems too hard to solve.
How does Recraft answer an unmet need?
Designers don’t want the output of image generators to differ from their expectations, so we need to give them more sophisticated control over the output.
Recraft focuses on giving its users control over the image creation, and the ability to iterate on the generated results.
To use AI image generation in design, it is necessary to support style-consistent image generation. To create a website or an app, you need several images in the same branded style. This challenge hasn’t been solved before by any other tools and we are working on providing this functionality to our users.
What’s in store for the future?
We have quite a lot coming up! To name a few highlights…
- High quality image generation with our new foundation model.
We are building our own foundation model, which will improve results for designs that involve anatomy. For example, generating means correct fingers, faces, and human figures. It will also provide higher quality details.. It will also have greater prompt consistency. And it will be possible to use more complex scene descriptions than it is now.
- Style consistent image generation – create a new style from an image or set of images.
We are working on providing the capability to create a style from a brand image or a set of brand images, so users can then use this style to generate more illustrations, icons or photos in the given style. This functionality will gradually improve with the future releases.
- Illustration and icon sets
We are working on offering the generation of illustration and image sets out of the box. It is important to have style-consistent imagery if you want to create an app, build a website, or support your brand with visuals. Image set generation will make it easier to reach this goal.
- Graphic design generation – generate branded graphic designs, like social media posts, ads and packaging from text given a style guide
It is important to not only generate illustration and photography, but also to build designs where images and text are combined in non-trivial ways. Graphic designs can be generated with AI. But right now, as far as I know, there is no company that does this in a high quality.
What one piece of advice would you give other founders or future founders?
To the future founders, I’d give the following advice:
If you want to start your company, go do it. There will always be things you think you can learn before, but the quickest way to learn it is to get your hands dirty. I waited before starting a company, because I thought I should learn first. But I would have learned it faster by doing it in practice. Also you don’t really know in advance what you have to learn, so your guess might be wrong
To the existing founders my advice would be:
Try to hire only the best people. Don’t compromise. If you manage to do so, the company will be a magnet to great talent in the future, and it will help you all the way from having a quick start, to growing the company in 5-10 years, when it will be the place where everyone wants to work, because it has the best people.
And finally, a more personal question! What’s your daily routine and the rules you’re living by at the moment?
I have a toddler at home, so my day always starts with preparing a lunch box for him and taking him to the nursery. After that I either go straight to the office, or do sport first, and then go to the office. Normally half of my work day is meetings – meeting investors, potential users, having UX interviews, talking to the team and to potential hires. Other time is spent on strategy and tactics, and doing a log of back office, testing the tool. I go home in the evening, help my son with his homework, and then put him to sleep, and go to sleep myself. I do need at least 8 hours of sleep every night, otherwise I’m not functioning.
Anna Veronika Dorogush is the Founder & CEO of Recraft.