From the NLM Tech Bulletin:
In March 2013, the National Library of Medicine (NLM) released a search-based Web service that provides access to the metadata and full-text OCR of all resources in the Digital Collections repository in Extensible Markup Language (XML) format. The Digital Collections repository includes more than 9,000 digitized monographs, films, and other resources the from NLM collections. Software developers can use this Web service to incorporate the historical resources into their own applications.
The Web service accepts keyword searches as HTTP requests and returns XML with metadata about and keyword-in-context snippets of relevant resources in ranked order. Each returned record includes up to 13 Dublin Core metadata fields and the full-text snippet. The resource’s permanent URL is included in the “dc:identifier” field. It may be used by the application to link users directly to the resource on the Digital Collections site, where users can view the entire digitized resource via integrated presentation software.
The web service is updated weekly. It is free to use and does not require registration or licensing.
Read the Complete Announcement
Direct to Web Service Documentation