About 1/3 of the 1M ebooks on OpenLibrary.org have full MARC records, 
and you can retrieve the record via the API. There is also a "secret" 
record format that returns not the full MARC for the hard copy (which is 
what the records represent because these are digitized books) but a 
record that has been modified to represent the ebook.

The MARC records for the hard copy follow the pattern:

https://archive.org/download/[archive identifier]/[archive 
identifier]_marc.[xml|mrc]

Download MARC XML 
https://archive.org/download/myantonia00cathrich/myantonia00cathrich_marc.xml 

Download MARC binary 
https://www.archive.org/download/myantonia00cathrich/myantonia00cathrich_meta.mrc 
<https://archive.org/download/myantonia00cathrich/myantonia00cathrich_meta.mrc> 



To get the one that represents the ebook, do:

https://archive.org/download/[archive identifier]/[archive 
identifier]_archive_marc.xml

https://archive.org/download/myantonia00cathrich/myantonia00cathrich_archive_marc.xml

This one has an 007, the 245 $h, and a few other things.

Tom Morris did some code that helps you search for books by author and 
title and retrieve a MARC record. I don't recall where his github 
archive is, but I'll find out and post it here. The code is open source. 
We used it for a project that added ebook records to a public library 
catalog.

You can also use the OPenLibrary API to select all open access ebooks. 
What I'd like to see is a way to create a list or bibliography in OL 
that then is imported into a program that will find MARC records for 
those books. The list function is still under development, though.

kc

On 8/18/14, 3:04 PM, Stuart Yeates wrote:
> There are a stack of great free ebook repositories available on the 
> web, things like https://unglue.it/ http://www.gutenberg.org/ 
> https://en.wikibooks.org/wiki/Main_Page http://www.gutenberg.net.au/ 
> https://www.smashwords.com/books/category/1/newest/0/free/any etc, etc
>
> What there doesn't appear to be, is high-quality AACR2 / RDA records 
> available for these. There are things like 
> https://ebooks.adelaide.edu.au/meta/pg/ which are elaborate dublin 
> core to MARC converters, but these lack standardisation of names, 
> authority control (people, entities, places, etc), interlinking, etc.
>
> It seems to me that quality metadata would greatly increase the value 
> / findability / use of these projects and thus their visibility and 
> available sources.
>
> Are there any projects working in this space already? Are there 
> suitable tools available?
>
> cheers
> stuart

-- 
Karen Coyle
[log in to unmask] http://kcoyle.net
m: +1-510-435-8234
skype: kcoylenet/+1-510-984-3600