The battle is heating up in the AI video space. While the promise of OpenAI's SORA has yet to be realized, new entrants such as KlingAI are really leading the way. One of the veterans in the space is RunwayML, a New York-based AI video pioneer that has just released the third generation of its already very impressive video product suite.
As the AI video field becomes increasingly competitive, the bar for quality and performance is being raised. A year ago, awkward, distorted video clips of actors eating spaghetti were acceptable; today, truly amazing, lifelike footage is being projected onto our screens. To compete, Runway will need to significantly improve not only the resolution but also the consistency and video fidelity.
Runway Review: Features
(Image courtesy of Runway)
It's important to note that Runway is not just a video generation tool. In fact, the platform is packed with a variety of tools to edit, customize, and enhance videos and images. There's text-to-image conversion and green screen background removal. There are also tools for image-to-video conversion, generative audio, upscaling, and frame interpolation. In total, there are 32 media tools to cover almost any media editing purpose you can think of.
There are five plans available, ranging from a free plan with 125 credits (non-renewable) for 25 watermarked images, to an unlimited plan for $95/month with 2250 Gen 3 image credits and the ability to train models on specific styles. These aren't the most generous plans on the market, especially compared to a growing number of international rivals, but for most people the standard plan with 625 Gen 3 image credits for $15/month is a good option to get started.
Five plans (including a one-time free plan) Advanced video editing works seamlessly with AI integration and offers project and team features targeted at enterprise/creative studio professionals
Runway Review: First Impressions
The home screen dashboard gives instant access to the tools most users will need on a daily basis – text to video, video to video, lip sync video and background removal seem to be very popular. The menu choices are enhanced by the ability to tag any option as a favorite, at which point it will appear at the top of the screen, making navigation even easier.
The business-oriented nature of the product is apparent in the prominent placement of the Assets and Workspaces sections in the sidebar. Here, you can organize and share all your work with folders and custom tags, making it easy to build a portfolio of business assets that can be quickly and easily reused. This is one of those features that often gets overlooked in new platforms, but is crucial if you have a team that works on company-wide material on a daily basis.
Moreover, Projects combines the best of video editing with the smarts of AI, giving team users more control over their workflow, which is important for business users too. Basic editing features revolve around traditional video timeline motifs, so they're not just completed with AI prompts. You can do some serious editing as you create your clips, including green screens, motion tracking, and filters. At that point, things start to get a little complicated.
This comprehensive editing suite is what sets Runway apart from its competitors, but this strength is also its weakness: to get the most out of the platform, users will undoubtedly need to take the time to study the extensive tutorials in the integrated help system, Runway Academy. The combination of sophisticated AI and advanced editing allows you to create almost anything you want or dream of. However, learning from scratch is no easy task.
Runway Review: In Use
(Image courtesy of Runway)
Many users, especially those on the lower pricing plans, will see Runway as just an AI video generator, and to be fair, you can get plenty of value from using it that way. At this level, you only use the prompt boxes and make basic edits to the results.
For example, I quickly set up a lip-sync video in about 5 minutes and the results were very impressive, although I don't think it would have been nearly as impressive if I'd used an avatar tool like HeyGen with the ElevenLabs voice – it's a much shorter and less cumbersome process.
RunwayML Lip Sync and Green Screen – 1080p – Watch on YouTube
To some extent, I think this is intentional. A quick look at the platform's user gallery underscores the company's desire to be a leader in the arts, not just a video editing service. This nod to Hollywood is apparent in nearly every video you see on Runway Studios' gallery page. The demo seems to say that while you can certainly make a cheesy corporate marketing video, they actually want you to use the company to create art that can win awards. It's a great differentiator, as Midjourney is intentionally appealing to the artistic image creators, rather than the marketing hackers of the world.
(Image courtesy of Runway)
So, with that in mind, does Runway offer what the mass market really needs? The answer is, in a loosely phrased way, yes. While the various tools offered are perhaps not best in class, together they provide the perfect mix of AI goodness that anyone can use when they need it. It didn't take long to take a lip-synced video clip, push it into a green screen editor, and completely replace the background. What would have taken a fair bit of work a few years ago now takes just a few minutes.
I tried out most of the tools on the platform and everything worked as expected: it looked really cool in slow motion, I used scene detection to split the video into scenes, I did a background remix for a mix effect, and restored black images to color. It all worked well, but some of the image generation was definitely off, especially the faces. I even created some fruity 3D textures to use as some of my imaginary game assets. But all the while I couldn't help but feel like I was missing the point.
(Image courtesy of Runway)
So Runway is clearly more than the sum of its individual tools. Unlike most other AI tools on the market today, Runway is designed to be used by true visual artists, not just enterprise users or consumers looking for a quick video hack to meet a deadline. Just as it takes time to learn how to get the most out of the platform, it also requires an “artist's eye” to combine the pieces into something beautiful and worthwhile.
The company's founders met at an art school in New York, and the company's platform has been used to edit Hollywood blockbusters, music videos for Kanye West and others, and TV shows like The Late Show. This isn't your grandma's AI text-to-video product. Setting aside the recent controversy over training models on YouTube material, consider the surreal fact that the company actually runs an annual AI film festival in New York and Los Angeles.
Runway Review: The Good, the Bad and the Ugly
It's important to remember that RunwayML is no Adobe Inc., a 40-year-old software company, but its Gen X tools have arguably revolutionized the world of video editing more than any other company in recent years, just 18 months after its first product launch. While Avid and Adobe hold the top spot in video editing among professional filmmakers, the depth and versatility of Gen 3 challenges them in terms of using AI smarts to deliver high-quality editing results.
That being said, however, this nascent platform still has a long way to go before it can grow into a true contender for the throne. For starters, getting top-class results at a professional level is still a fairly difficult process. The company's heritage as a co-developer of stable diffusion image generation technology is both a strength and a weakness. It's clear that the focus on delivering innovative video has put image generation development work in the shadows, which shows up in the inconsistency and low fidelity seen in many of the Runway Gen 3 results. But since Gen 3 is still in alpha, we should probably forgive it.
But what I really need help with at this stage is the interface before it's too late. Somehow, the user experience felt like walking around a messy teenager's bedroom – things are scattered all over the place with no discernible order or pattern. For example, why are there two separate heading menu options for Generative Audio and Lip Sync Video when they do the same thing?
And why don't we have two functional interfaces that look and feel consistent? Each component of the platform is confusingly different when the functionality and user interface could easily be made more consistent. And whose bright idea was it to bury the primary video editing functionality under a (More) button in the sidebar? Ouch!
Runway Review: Conclusion
(Image courtesy of Runway)
In fairness to RunwayML, UX and UI are often the last items tackled in the cutthroat race to achieve product-market fit. Building features, fixing bugs, and responding to market demands often consumes all the air in the room and barely addresses the usability needs of your poor customers – just ask eBay. But the honeymoon period with AI applications is rapidly coming to an end, and if companies don't address some of the key issues, it's very likely that someone else will come along and steal the market.
My experience with the platform, limited by my total lack of artistic skill, has been enough to prove that the product has a bright future in every respect. It's not sophisticated, but it's clearly very powerful and has all the elements needed to build a product that rivals anything else on the market for multimedia creation and editing. The seamless AI integration is a definite bonus.
Ultimately, I couldn't help but feel defeated by the toolkit. My mundane efforts at creating an AI video clip were just scratching the surface of Runway's capabilities. Sure, there are AI image and video manipulation tasks that can be handled in the same way that a Ferrari can help you with your grocery shopping at Walmart. But for professional video creators who need to think outside the box and showcase their talents at the highest level, I think few can compare to the Runway Gen X tools on offer here.
