Hacker News new | ask | show | jobs
Show HN: Convert blog articles to high-quality audio (websitevoice.com)
146 points by tiburon 2720 days ago
20 comments

If this is not a wake up call for me, I don't know what it should be. I had this same exact idea for 9 months now but never moved to execute. Reading through site it even has the same wording as in my head.

As a developer I wanted to build something like this and it looks easy(without AI part). But following the startup gurus and podcasts I figured that idea should be validated first and to find actual customers is more important than building technology. Here started troubles. I have no skills in marketing and when diving deeper it looked cumbersome for me. I spend free credits on AdSense but it only brought 8 emails for empty newsletter.

I'm interested did you do the research on user base? What would be a market for this?

Where are you planning to grow the site? Do you have monetization plan?

p.s. found a link in comments for read2me.online, and just from demo they sound really good for ears

Likewise, I spent a day building a proof of concept but lost interest after.
Hello, Very interesting. I've been using Edge's built-in voice reading a lot more recently for the exact case you mention on your homepage: listening while working. I have two questions for you after reading your TOS and Privacy Policy. I was expecting to see "you" scraping data from each site or otherwise monetizing the service, but nothing stood out. Either your TOS/PP is incorrect, or you don't have an obvious income stream for the service you are rendering.

So the questions: Who are you? and How are you doing this for free?

Thank you,

Good to see others checking TOS/PP too, and asking these questions. Especially as there is absolutely no transparency. Who is the company, or some people behind this, contact email? The TOS says 'use at own risk, we are not liable for anything' and also mentions 'Refund policy' (?).

The domain is namecheap, no more info. And the HN user tiburon is a new account.

There are 4 trackers on the site. Besides GTM (mentioned in PP) there is twitter-ads.

Do you have more info, tiburon?

currently it’s free and there will be some paid features in the future, the site is preparing for that.
Yea, I read the indiehackers thread and understand you currently focus mostly on the technical side.

I assume you'll have the business side well thought out too. My feedback would be to be as transparent as possible, once adding commercial services, and turn that into a true USP.

thanks for the input. We have built several products before and we follow the same strategy almost with every new product.
In light of the various concerns, it would be helpful (read: to many here essential) if you elaborated more. What other products? Is there a contact email? Etc.
why it really matters on who I am? I'm a developer with passion to help people listen to podcasts and the idea was simple. How can I offer it for free? I do a lot of work to make it happen, there will be a premium version with more features which are being cooked with the help of many developers who loved the idea too.
You know ten years ago I would have 100% agreed with you. In fact I would have been outraged as to why anyone would want to know.

And yet today, I consider the provenance of a product to be more or less my No. 1 consideration in whether I will use it or not.

We live in a world where brands and products are established, then monetized and eventually their userbase becomes just a trading chip that passes from hand to hand. We're giving access or money or trusting systems and we really have no control over the transfer of that trust, data or money.

So unfortunately now, yes it really matters a lot. You're asking me to add code to a website of mine: I want your address, your email, your phone number...I need to know who you really are.

Ah, I just missed this comment when sending my reply.

Many people indeed do not care, and just use any cool service they encounter (and yours is cool, don't get me wrong).

An increasing body of people (including me) is more wary of just using any stuff that is out there on Wild West internet. And for good reasons.

In your comment you provide information that would be very valuable to have on your site.

E.g. 'I am a passionate developer and intend to provide Premium services eventually. The basic services you find here will continue to be free.'

Given the purpose “to help people listen to podcasts”, I’m a bit puzzled about which end you’re tackling it from. Does it not make more sense to work on the user’s side, on browser software, where you can make all sites work, rather than helping the odd site here or there to support this? When approached from the website side, I don’t expect many to adopt it.

(On the browser side, I am aware that various browsers have speech controls hooked up, most commonly inside a Reader Mode—to say nothing of screen readers, which have a different target market. There’s also the Web Speech API which could be used to replicate WebsiteVoice’s functionality using whatever voices are offered by the user’s device. Its quality will be heavily platform-dependent.)

It matters because one of your potential customers thinks it matters. And presumably because services without a business model don't have a long future.
timClicks that is not true, there are many products out there that do not rely on capitals, It does not matter who I am personally. The service is free of charge until we cannot make it free of charge anymore for whatever reason. We are focusing now on user's feedbacks who already added the widget. you can find more answers on our business model there. https://www.indiehackers.com/forum/finished-mvp-to-turn-blog...
Out of curiosity, are you not some variation of a programmer or engineer like the majority of people here? Asking because I wasn't able to combine my work with anything demanding mode attention than music.

(However, I must say that audiobooks go splendidly with manual work that doesn't require much thinking.)

I think there are many services similar for free, they most probably have their own TTS
There are some UI issues from what I noticed:

At small breakpoints, the UI (especially the "try it now" button) can overlap some text. Sizing also seems off.

Another issue I noticed is the play button itself, the entire circle area is clickable, however only the play/pause icon can actually be clicked.

There are also numerous accessibility issues. If you're going to advertise this as something to improve accessibility, it might be good to actually concentrate on that for your landing page.

I do hope the UI can be improved, as the product looks solid.

I'm not here to pitch the idea to make people like it, if you do not like it, be constructive on a given problem. To get a feeling more on how it feels on a blog here is an example: https://websitevoice.com/blog/text-to-speech-dyslexia-and-re...

if the solution does not appeal to you, it surely appeals to others.

Couldn't listen to it for more than 20 seconds, got quite robotic and monotone for me. This would be like flash and auto play videos, cool at first then most annoying thing ever created.
monotone? hahaha you should try the default text to speech from the operating system to see what is monotone
@tiburon thanks for the feature, i truly love it. i've added it to my blog with the JS snippet and it plays now on all the articles! https://jaywad.com/blogs/tips-for-a-productive-morning-routi...

I would wish to be able to customise the play button though.

https://read2me.online Another similar service
Pricing looks expensive, but quality is top notch. Do you know which TTS engine it is?
This looks -- and sounds! -- awesome. I'm not sure if this can compete with Schema's 'Speakable' [1] property (for smart speakers), but for mobile/web browsing this looks amazing.

Looking forward to seeing new developments and also the pricing plan eventually.

1: https://developers.google.com/search/docs/data-types/speakab...

We created this free web app to convert articles to high-quality audio. Our mission is to help people with learning disabilities, those who prefer listening over reading, and who are always on the go to enjoy consuming your content online. Please give it a try and let us now what you think. Your feedback is highly appreciated :)
Since it's free, I wonder why don't you make it available as a browser extension instead of asking publishers and content creators to manually embed your service.
I'm working on a project that overlaps slightly (only with this feature). Would you be open to offering programmatic access in the future via an API?.
@jamil7 what are you working on? I wrote here that I've had pretty similar idea, wonder what's your take on that?
@mindwork DM me if you feel like it be cool to hear what you're working on.
The audio is anything but high quality!

Please use better compression settings.

I added it to an old unused part of my website.

https://www.folj.com/lateral-solutions/

Apparently my javascript doesn't play nicely with programmatically injected elements. The Show All Solutions button still works.

good, at least you powered your blog with voice. happy to see it useful for you.
Just some hopefully constructive criticism, I found the voice used on the "Smart Natural Voice" section to be the most robotic sounding of them all. Specifically, the "Say goodbye to robotic voice, this is nothing like you ever heard before." line sounded very robotic and a bit awkward.

I'm not sure if maybe the content of the copy is a little clunky or if it's the voice itself, but that section sounded worse than the other ones, IMO. If possible, I would think about toying with that. The girl's voice in the "Voice Pitch and Speed Control" section sounded far more natural and impressive.

Can I use this to convert other articles? Could I pay to get 32kbit Opus?
Do you have plans to publish a versioned release with support for Subresource Integrity?

Also, is there somewhere that the following is documented?

- Required origins for each type of content so that this can be compatible with a Content-Security-Policy

- Required use of browser features so that this can be made compatible with a Feature-Policy

- Commitments from yourself that the lists of required origins/features will not be added to without prior notice, in order to give time for blog operators to update their headers accordingly

Thanks

I’ve wondered for a while with NPR doesn’t have both audio and text of all of their programs.

Sometimes I can listen, but not read (driving). Sometimes, I can read but not listen (work, or just when I’m without headphones). I realize they want high quality voice, but if at the moment all they have is text that’s ok with me. Not having both of these often means I just skip that article and never come back to it

radio ambulante (NPR's spanish podcast) has full transcripts in english and spanish of all their content. The quality of the journalism and topics are outstanding as well.

Super useful for learning Spanish.

Amazon Polly [1] and Google Text to Speech [2] are unfortunately far superior.

[1] https://aws.amazon.com/polly/

[2] https://cloud.google.com/text-to-speech/

I have to agree, plus they’re free for what most users would need.

Also noteworthy is the official Amazon Polly for WordPress plugin - https://wordpress.org/plugins/amazon-polly/

Don't have time to test it now. How does it sound? Is it better than the robotic text to speech voice we all know? Wonder if you can run this offline so you can save the files when you build your static site and no scripts are required on the frontend.
It's certainly better than most TTS out there. We are developing AI in-house to make it as humanly sound as possible. There's always improvement need to be made, but we're getting there :) Your latter suggestion is definitely worth something to look at. Let me know once you've tested it out. We highly appreciate any feedback. Cheers!
I'm curious how the economics work out with building your own TTS vs. just paying Google to do it for you. Of course Google's TTS API costs money, but I assume that you also had to pay for a dataset of transcribed speech. After how many hours of audio is the breakeven point?
The actual text to voice synthesis isn't acutally bad, but the audio quality is terrible. The audio artifacts are so bad I could barely listen to the demo snippets. It's too bad they ruined a quite good product by skimping on the most important quality, audio quality.

I wonder how OP missed this? Audio quality has to be flawless when using headphones for extended times. It's not like you need a lot of bits for audio, so it's not something you save on.

Please fix!

Sounds good on my speakers
Headphones.
It's fine on my phone and laptop. I think you are the one with the issue.

I would delete my comments if I were you because you look like a jerk mentioning it 3 times in a post with 8 comments total. And it's not even true.

the audio sounds good to me too even with headphones, are you sure you haven't got broken headphones?
Laurel or Yanny test?
it sounds great for me too. you must have headphone issues.
You should showcase your product better.

Add a text input box so that prospective users can try it themselves.

Typo: "We use Artificial Intelligence and Machine Learning to constantly improve our voice algorithms to make your website text-to-speech is as realistic as possible."
Thanks PapaSlug, i like your username. The Typo is fixed now.
Constructive criticism , I could not find any samples of the audio on your site.

Recommendation, add the plugin to your own homepage for potential users to test it out before having to signup.

thanks for the input:

there is a demo button here on this page where you enable your website. You can even test the feature without signup.

https://websitevoice.com/convert-text-to-audio-free

Apologies, looks like your widget domain is being blocked by my work and was preventing the demo button from doing anything.
There is a large 'try it now' button above the fold.
What I see (image: https://postimg.cc/3yg6p19w) is a "try it now" arrow, but no button, anywhere above the fold.
What's the catch? How are you providing this for free.
according to what he said on another thread it will be free and it will have some paid features on top
I've also tried to see what was the catch but seems like they just want to make it for free, maybe to help people with disabilities. Or maybe the add payment features in the future
as long as its free, its good. lovely idea! keep it free please.
Is there a demo of the UK voice anywhere? I could only find the American one.

Very impressed with the audio quality. What's the revenue model?

>Very impressed with the audio quality

To me it sounds not much better than Google's standard wavenet TTS (seen here: https://cloud.google.com/text-to-speech/) which is admittedly pretty state of the art but nowhere as good as actual human dictation.

I use Read Aloud Firefox extension to listen to blog posts.