Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: Instantly listen to any URL

per.quest

Post reply on HN
Print view View original

81–89 of 89 posts

Re: Show HN: Instantly listen to any URL

#81

Post by hopesthoughts »

neolog wrote: post #54
busymom0 wrote: post #51

Earlier quoted context omitted.

Not sure about OP but I just implemented this in my Hacker News android client (thanks for the idea OP). This is how I implemented it. I had already achieved article to "reader mode" extraction by heavily customizing the Kotlin port of Mozilla‘s Readability: https://github.com/dankito/Readability4J From the extracted "reader mode" text, I had to do further extraction to get rid of things like image captions, author n…

I really want one of these that can follow RSS/Atom feeds.

Well just download NVDA, and run it through you're RSS reader. It will read RSS feeds for you. I don't think that's what you're getting at, but it will do that, just as it does other things.

Re: Show HN: Instantly listen to any URL

#82

Post by hopesthoughts »

dreadlordbone wrote: post #48
croes wrote: post #12

Earlier quoted context omitted.

A download option could lead to copyright violation claims.

That makes zero sense. It is already being downloaded to the browser.

Also, if you don't do anything with it afterwards, what kind of claims does anyone have?

Re: Show HN: Instantly listen to any URL

#83

Post by hopesthoughts »

hemloc_io wrote: post #23

Very cool! I've made a small prototype of something like this before, but what's always gotten me is the limitations of text to speech. Hopefully some point soon we will have options for more natural speech.

What people who use TTS on a regular basis usually want is something that gets the job done quickly. This is why natural-sounding TTS isn't generally a good thing. That being said, there are some people who like it. I'm definitely not one of those people.

Re: Show HN: Instantly listen to any URL

#84

Post by hopesthoughts »

sntran wrote: post #16

This is timing. I was just looking for a solution for a friend who has very bad eyesight due to accident. They want to be able to listen to news articles and such. It would be nice if we can choose languages other than English.

They may want to look into using a screen reader to avoid eye strain, or headaches depending on what their condition is. Everything has one built-in at this point. Windows has Narrator, the Mac and iPhone both have VoiceOver, and Android uses Talkback. For something extra, and more advanced, I definitely recommend NVDA. https://www.nvaccess.org/

Re: Show HN: Instantly listen to any URL

#87

Post by remram »

soheil wrote: post #56
eat_veggies wrote: post #36

Is this somehow using the waybackmachine to scrape pages? When I try some of the classic ssrf paths like file:/// or http://localhost:8080 , I get audio about the internet archive. When I put in https://localhost:443 , I get audio for the default nginx page, which matches what the internet archive has for that url: https://web.archive.org/web/20210620003533/http://localhost/ But putting http://localhost returns the a…

Added a check against localhost thanks for pointing that out.

You will not win against SSRF using a blocklist of names. You need a comprehensive solution that is designed for this and will check the end address after redirections, DNS resolutions, ... have been applied in whatever library you use for HTTP. An example is advocate for Python https://github.com/JordanMilne/Advocate

Re: Show HN: Instantly listen to any URL

#89

Post by call-me-al »

axegon_ wrote: post #47
yunusabd wrote: post #46

Earlier quoted context omitted.

Same here, we built a scraper to (more or less) intelligently extract text from websites, I think adding TTS to it would be trivial. I'm guessing the hardware requirements are for the TTS engine? Could the Web Speech API be used to generate the audio locally to bring down costs? Even though you'd lose support for IE and FF.

Exactly what I did, simple web interface I run inside rambox along with all chats and email clients and whatnot, paste a url and it scrapes it and generates the audio to be played in browser(rambox in this case). OK, I do keep a copy of the audio and the text and some other metadata but that's completely optional. As far as tts, while the cloud speech is unmatched, it's still proprietary and not free. However there a…

that sample sounds very convincing! if you don't mind me asking, what TTS library did you use to generate it?
Post reply on HN