Runway builds tools for human imagination. Using next-generation AI models, we‘re augmenting the creative process with a focus on control, expressiveness, and flexibility.
With our iOS app, you’re now able to harness the full power of generative video from anywhere with Gen-3 Alpha. Our latest model allows you to generate hyper-realistic and controllable videos. All with a simple text prompt.
The way it works is simple:
- Type what you want to see
- Tap generate
- Share your video
Updates and new features are being rolled out regularly, so you'll always have access to the latest. If you already have a desktop account, you can also transfer assets seamlessly from your phone to your computer.
So far, I have tested and ultimately been disappointed by a lot of AI art generation apps, this is, in my opinion, a game changer.
At the moment I’m on a free plan which is very generous with the credits, to buy more credits or a monthly payment, even annual, it may seem very expensive, there is an enormous amount of computing power to create these stunning videos, so I totally get the price, it’s fair.
So far on a single prompt, with camera controls, style of output, no starting image at first, I was blown away how it created exactly what I envisioned.
I’m not saying that is always going to the case but just don’t be ridiculous with your expectations .
I would imagine starting of with an image you provide, would create superb results, for me I like the exploration and surprise of the outcome.
Great feature, is the ability to continue the video, lengthen the video, if you like the outcome, to me this is a great option.
Clarity, resolution is great.
Absolutely recommend a try of this app.
Small school girl extreme close up
I wanted an extreme close up, but the generated video shows medium shot.
Huge Potential
So far, the best I've found for AI video generation. The prompt to animation is OK but not great. The image to animation is very good already but not perfect...it often stumbles on movement, esp walking. But for limited use cases (it only generates 4 seconds of animation and it does certain shots much better than others) it's really really fun. I think maybe overpriced for what it does right now...I'd say 30% of the time it'll give something usable. It asks for feedback on every single shot which gets annoying, you are effectively paying to train their model.
But. It's fun. It's addictive. And when paired with something like Midjourney to generate images, the results can be remarkably good. Looking at the improvements in other areas of AI in the past year, I can't wait to watch this improve over time.
Wow
The app is great just learn to prompt well and it’s easy
Works with prompted images
Works okay with Gen-2 when you submit a pre-generated image, creates a basic movement, but still takes a couple of edits to create the result you want. Doesn’t want work with a freestyle text prompt as too much warping and literal interpretation.
Can’t log in - won’t send email
FIX YOUR APP. Would love to be able to use this app or the desktop version but after sign in, neither app nor desktop recognised the password or the username. Email password re-sets don’t send. Nothing comes to my email. Be warned, for now new users will just be locked out.
Fantastic
Everything it says on the tin mate
None of the outputs can be used for anything
The quality is really poor, you might get a preview of a half decent looking image, but the moment it is turned into a video it starts to warp, losing any consistency, turning into bizarre surreal weirdness.
This is fine as a curiosity, and I fully understand that text to video AI needs users to develop, but paying for this app right now, is basically funding Runway.ml training algorithm development.
Don’t be fooled by the previews and trailers they show to promote this tech, it’s nowhere near usable, and they shouldn’t be charging people subscriptions let alone tokens for outputs that don’t come close to the initial preview.
Give it a few years and I’m sure RunwayML will fix many of the issues, it’s just not fair on the people who are paying to test it in such an early state.
Might be good but as it doesn't work who knows
Well I don't call this a revolution in AI VIDEO generation, mainly because it works so rarely. When it does it's meh at best and the idea of paying £15.99 a month for such a hit and miss app with iffy generation ability is laughable. Hopefully they can either get their act together or someone else will provide a product which can do this properly. I'm not unrealistic and appreciate the limitations at present but this is laughably badly done. Such a shame!
Don’t waste any money on this
The current generative ability is extremely poor. Poor matching to prompt, poor natural movement, and poor overall render. They have moved to a paid service much too early. And they’re basically asking users to pay to test their awful ai model and give them free feedback to improve it. A total waste of time.