24 comments

[ 2.7 ms ] story [ 70.0 ms ] thread
Very Cool! I like the simple no-nonsense UI. What is your goal with the project?
Thanks! :) No big goals with this. Thought it'd be a useful tool, especially considering how big and bloated web sites are getting these days. So we're going to try to keep it around as it is. It's a simple wrapper around our Full-Text RSS product, which, as of the latest release, can output plain text like this. Demo of that available here: http://95.216.144.183
Nice!

Suggestion: allow larger line widths to be specified (for fewer, longer lines in large paragraphs).

Thanks! I'll try to add that in the next update.
Hey this looks really cool. I think the typography could be improved though, check out http://bettermotherfuckingwebsite.com/ for some inspiration/guidelines on how to accomplish that.
Thanks! :) The output is literally plain text (Content-Type: text/plain).

But, I realised it's possible to send a stylesheet in the HTTP header. Chrome seems to ignore it, but Firefox uses it:

  Link: <https://txtify.it/dark.css>; rel="stylesheet"
Very cool! Could you also add viewport css so that it looks better on mobile Firefox? (it already looks nice)
Thanks! Tried adding this to dark.css

  @viewport {
    width: device-width, initial-scale=1;
  }
But didn't work in mobile Firefox.

Seems Firefox and Chrome on mobile handle plain text differently. Firefox shows the text quite small (smaller than if it tried to fit it to device width). Chrome on Android, on the other hand, doesn't seem to try to fit the content at all, showing it at fixed scale which causes additional wrapping to the already 70-char wrapped text we output.

Happy to experiment more with CSS, but ultimately this is simply supposed to be plain text output, so optimal readability isn't really the goal here - although would be nice to improve it. :)

I’ve frequently needed to access a plain text version of a site in a bash script or from the command line. For this, I use the command

    w3m -dump <url>
which dumps the website’s text content to stdout.
One of the things we do with txtify.it is try to find the content block in the HTML first, then convert only that from HTML to plain text. So in situations where the input URL is a news article or blog post, the output should be cleaner.
Where is your privacy policy?
We'll have one up soon. But essentially no user accounts, and no logging on our servers for this service. Requested articles are cached for a few hours then deleted.
I count this as a proper privacy statement already :)
Just a little feedback, there's a small problem with The Economist. It just includes "get our daily newsletter" thingy. Also cool project!
Have you got a URL I can try? This one worked:

https://txtify.it/www.economist.com/news/finance-and-economi...

Here is it: https://txtify.it/https://www.economist.com/news/leaders/217...

It also shows up in your example as well. Maybe you don't get bothered by it? :)

Please note: the bug is inconsistent, it pops out inconsistently.

Here's the part: "GET OUR DAILY NEWSLETTER Upgrade your inbox and get our Daily Dispatch and Editor&#x27;s Picks."

I thought you weren't getting the content text at all. :)

Should be fixed now. We now strip that out: https://github.com/fivefilters/ftr-site-config/blob/master/e...

But I don't consider that a big problem - lots of sites include things like that in their content. It's annoying, but no easy way to identify them all accurately and remove them without custom rules.

Oh, it's certainly not a big problem, I just wanted to let you know :)

liked the project by the way!

Thanks! :)
The example should rotate between what people have previously converted. Right now it is static.
That would bother people. Say you convert something that's private or not public yet. Not sure you'd want the next person to see it. I don't like it when sites do it.
Am i right in thinking the source code will not be available for this?
Any particular aspect of the code you're interested in?

The article detection, extraction and conversion happens with our Full-Text RSS product which we sell (see link for demo in another comment). So the code here is mostly sending a HTTP request to that and asking for plain text output. The HTML to text conversion in Full-Text RSS uses the Html2Text library found here https://github.com/mtibben/html2text

Cool project!Is it open source?