This sounds like it would absolutely ruin the internet as a resource for how-to videos. When anyone can use massive amounts of data scraping and underpaid labor to make extremely low quality videos and spam youtube with them, actually half decent how-to videos will get completely lost. Not to mention, you can just put something completely wrong on without making it immediately obvious.
Imagine Roel Van de Paar[1] except anyone can easily do it and the quality isn't so bad that it gets lost in the algorithm.
Then consider the internet ruined, auto generated how-to videos exist on Youtube and elsewhere (probably in much greater numbers than easily spring to mind). I see them from time to time in the long tail. But I suspect mostly the search algorithm deals with spam reasonably.
Cheap low-effort AI-narrated training videos are exactly the problem alexb_ is referring to. WowTo seems like a good tool for that, so if YouTube spam isn't what you're intending, how would you control what people use it for?
Imagine you are a plumber that needs to onboard new people at your work and they need to know how to do a plumbing thing. Turn over is pretty high and teaching each and everyone of them yourself takes a lot of time - so you make a video.
Fine idea, but I believe the better issue to tackle here is:
- how do I easily create and display how-to guides?
Maybe it's just me, but how-to videos are not as easily consumable as a simple step-by-step guide on a single page. All tools, I've tried for creating the latter, suck. Any ideas for that? :)
Oh but it’s not a fine idea. It’s expanding the scope of our problems. I have ADHD … how to videos are fucking poison.
I first noticed it with minecraft redstone machine explanations and tutorials. They started out as nice elaborate step by step guides with pictures… then they became videos full of calls to action from Patreon and twitch.tv and discord and their various minecraft servers…
Now I see it fucking everywhere and I’m sick of it. I want you to walk me through a process so I can follow it at my own speed. I don’t want your artificially slowed down video recipe with 6 minutes of filler life story shit to help you get that second YouTube advert or your sped up and fast cut edited tick-tock reposted to YouTube with on screeen clickable call to action pop ups kindly enabled by YouTube in their latest “crack cocaine for users attention” update and most fucking important of all… I do not want someone making it easier to produce this sort of shit for technical content that is hard enough to understand as it is from probably already average-to-poor quality documentation without getting a shitty AI powered video recap!
I hate having to say this but as much as I feel like an old fart on the lawn yelling at children to get off it… something something eternal September something something… it legitimately needs saying. Video is not universally the best way to share information and the fact that YouTube, et. al. will host your video content for free so you don’t have to try and work out how to make a web page, so you can focus on building your brand or whatever motivates the content in the first place … does not make it appropriate that we are sliding gradually into an idiocracy style world where nothing ever gets ducking written down properly.
And it just you know… makes my life harder and it’s very hard to be sympathetic to people making my life harder as their business model.
Edit: In case anyone thinks I’m being hyperbolic. I have bought at least dozen physical items in the last year where the only written instructions were legally mandatory safety info and how to plug it in. Everything else was QR code to their YouTube clip or companies YouTube channel…
Reveal.js is low overhead for this use case - create a stack of vertical slides (overhead is typically just an opening and closing html tag for each section of your guide) and use the javascript api to toggle overview mode on when the page loads (1 line overhead plus open and closing script tags).
I know you should typically use reveal for presentations but the overview mode and print to pdf feature are pretty usable for guides.
I like it primarily for the low overhead, but it has decent support for low-lighting irrelevant lines when presenting code (my typical use case) and i can use markdown for text and because it’s pure text in git, i find it easy to share with others when there’s collaboration needed.
"Do you often find yourself explaining about the documents to your clients or team? Now, let the AI do this for you with WowTo."
This doesn't sound native. I would change it to "Do you often find yourself explaining documents to your clients or team? Now, let the WowTo AI do the explaining for you."
Yeah I understand we will update it soon, but we do have a lot of samples created with WowTo.ai in the website, also our help center (https://app.wowto.ai/help) is a multi-lingual video knowledgebase.
We can but consider the super low attention span visitor who bounces because it is too hard to find the obvious thing you would expect/ask when a company is offering an AI video service.
I played the example video ("how to change language in Slack"). I liked the result as a video, but I kept thinking that I'd like it even more as a regular article.
If the input is photo and text then I think I'd prefer to consume it by just looking at photos and text. What's the value of generating a video out of them?
For context, I do watch how to videos too, but I do that for actions that are harder to describe with just text and pictures; like learning an instrument, whittling, skateboarding, etc. It's very likely that I'm not the target audience.
Yep, if I'm trying to find out how to do something I can do right now then I'd rather have an article. Otherwise I'd be pausing and rewinding a video multiple times.
If I'm just generally interested and it's not easy for me to do at that moment, then I'd rather have a video.
Since it's all generated I reckon an article and accompanying video would be a good product.
We are eager to announce to the HN community the public release of "WowTo" - software for creating How-To videos.
Pain point:
Creating How-To videos which can be tutorials, guides, step-by-step walkthroughs or video slideshows are difficult to create with current video editors in the market. Because they target a wide base so either they lack certain much needed features for creating How-tos or becomes complex to edit the video with their timelines.
Solution:
This presented us an opportunity, to address this problem with WowTo. It has a video editor that is specially created for making HowTo videos. Infact our first beta testers are elementary children who were comfortable using our product without any assistance or documentation.
Here is a list of features that makes WowTo - Wow!:
WowTo makes it easy to create howto videos with its natural AI voiceover feature, intuitive editor and multi-lingual support.
* No technical or creative knowledge required for creating videos.
* Select from 300+ AI voices with different languages
* Update videos for voice or content in just a snap
* With WowTo, let viewers listen in their native language instead of subtitles
You can also create HowTo videos from a variety of sources:
* Turn PDF document to a HowTo video
* Make a HowTo video from powerpoint in a single click and let your presentation do all the talking
* Use our Chrome extension to capture workflow steps as a HowTo video
* Have a screen-recording? Simply import it into WowTo and magically convert your voice to script
For more details visit our just released website: https://wowto.ai
Feel free to signup - there is a 7-day no credit card trial.
How about you provide a service where gag voice artists just dub over the AI generated voices so that the timing of the video stays the same but the voice quality gets immensely better?
I really like the idea, but watching some demos at the bottom of the page, I did not like the generated voices. They are unpleasant to hear, and I constantly compare them to crappy YouTube tutorial that have similar TTS voices.
I got it, I am sure it cannot be compared with human professional artist. However AI is improving and definitely it keeps getting better. And with WowTo you have many advantages:
1. If you do not like the tone of a speaker, there are several voices to choose from and also dialects
2. And you can instantly connect with a global audience with its multi-lingual capability.
3. Reduce your voice-over costs significantly for using with videos that does not focus on voice modulation but rather conveying the information.
I completely agree with this. My immediate advice because this tool is definitely more likely to be used by more people who aren't programmers, focus on quality. These types of users will have a higher bar for quality in this specific aspect. But also right now the quality of the TTS voices is too low in general imo.
Less technical users get more frustrated when their tooling can't produce things that meet their expectations. And moreso in my experience because they don't have the skillset in order to modify their tooling so it can meet their expectations.
> But also right now the quality of the TTS voices is too low in general imo.
Most youtube shorts (and I'd wager tiktoks, but I have no idea) have this incredibly irritating generic lady voice in the background. It's definitely artificial, every video has the exact same diction and tone. Clearly this is good enough for most people, so it can be done.
I really hope the quality improves before I'm forced to engage with such content. It's an absolute deal breaker for me. I can't reach for the back button quickly enough. I'm sure I'm far from alone in this.
Interesting product. It's not targeted at me specifically, but I could see it getting a lot of adoption. If this also had a programmatic API for video generation I would have some personal use cases for it! But I don't mean to say it should be an API/programmatic focused product because there are so many people who can't be served by tooling like that due to lack of programming knowledge.
Nice project, congratulations. I think there's definitely a use case, and for those crying 'Youtube spam', it's going to get expensive quite quickly if that's your application. One thing I would say though, is that your pricing could be improved. I would take a long hard look at Loom's pricing. It's designed to encourage both use and virality. Which are both essential to scale up. Don't limit the free trial to 7 days, and increase the number of free videos and time limit. But restrict distribution and make the watermark the motivator to shift to a paid plan. Lots of things you can play with to test adoption and referral.
Thank you! It looks great!
Just to understand the plans, p. example in "Plus" are you limiting the number of projects per month to 10 with each having a maximum of 10 minutes?
So, with 10$ in a month I can create a maximum of 10 videos where each cannot have more than 10 minutes?
In the following month can I create another new 10 videos?
Do I fully own the copyrights of the generated videos? To publish them anywhere?
First of all you own all rights to the video you create.
Regarding pricing, the pricing is based on the total number of HowTo projects in your account. The pricing is flexible , You can always upgrade or downgrade the project count. Hope that clarifies.
There was a very nice software called studio1 (defunct now i think) which allowed you to write phantomjs scripts (using a chrome extension) to create how to videos like this.
So instead of an image slideshow it was a very smooth video where it looked as if it was an actual screen recording and much easier to follow. Might be worth investigating as a feature upgrade, since that site shutdown.
If this is as easy as it looks, I can almost imagine creating personalized videos to help individual customers in tech support. Like Windows Steps recorder but with a publishable-quality output.
This look interesting. I'll need to test it out. I'm a, little it puzzled what does subtitle support means. I guess it is possible to embed them directly into video.
Thanks for your feedback. It automatically generates sub-titles in VTT format.
Or you can our own video player to display sub-titles in multiple languages for a video.
Embedding subtitles directly into the video stream is a bit more complicated. Interestingly a lot of video creation platforms don't support that as a feature even though it's useful in some cases (E.G. videos on instagram/tiktok/etc.) I actually had to write my own service on top of ffmpeg to do this shrug
Imagine Roel Van de Paar[1] except anyone can easily do it and the quality isn't so bad that it gets lost in the algorithm.
[1] https://www.youtube.com/c/roelvandepaar