No, it uses a custom scraper written by Marco. I've found that both of them work better on some pages and worse on others.

(However, I do believe the "Safari Reader" feature of iOS 5 and Safari 5 does use Readability to extract the contents.)

The "Reader" button appears to be a fork of Readability (http://downloadsquad.switched.com/2010/06/08/think-safari-re...)

Also of note, the original Readability was actually "partly inspired by" instapaper (http://lab.arc90.com/2009/03/02/readability/)

So I looked into all of these some time ago for an app that I use myself. The Safari fork of Readability is the best client implementation, while the best server-side implementation (by a mile) is Diffbot.

Out of the 50 sources I had (regular RSS updates, etc.) InstaPaper failed on far too many, either stripping important content or pictures and captions. Readability was only a bit better

I glued together an app that will pull in RSS, send each URL to diffbot, and then collate the output into a reader that I use. Never thought of releasing it as an app, I just use it myself.

Thanks for introducing me to Diffbot.

From your comment I expected it to be an open source implementation, but a high-quality SaaS with 50,000 API calls/mo on the free plan is still pretty damn good.

Readability's code is a little bit old (2009), but seems no other better (opensource) alternatives...

And thanks for suggesting Diffbot, will try later.

C implementation and Python bindings, opensourced one hour ago https://github.com/fiam/readable

Maybe the name is generic enough, but I'm wondering if you're aware of Readable[1].

[1]: http://readable.tastefulwords.com/