POST request sends data to a server in order to update/create a resource. POST request has similarities with GET request, except it contains a body. That data sent to the server is stored in the request body of the HTTP request. POST requests are not cached and cannot be bookmarked. They do not remain in the browser history.

POST is somehow safer than GET because the parameters are not stored in browser history or in web server logs and because data sent is not part of the URL.

Those for which true confidentiality comes with no compromise, search engines providing a cloud storage service regarding browser's settings is definitively not an option. The dully auto-stated, world's most private search engine has by default the right settings to deserve that appellation. Then POST method is coherently as default. Since any government surveillance programs that were begun with good intentions unaccountably have become tools for abuse, it even provides the ability to opt for the use of EU-servers, instead of US' ones, in order for EU users to avoid the mass Internet surveillance program "PRISM".
Despite the cookie StartPage sets to save settings is non-identifiable and therefore safe to accept, cookies in general may pose a threat to user's privacy.

For those users like me that do not accept cookies, an option is provided to generate an URL in an obfuscated form, (model: 'https://www.startpage.com/do/mypage.pl?prf=5ca7c2adc2e8c17819jk4eed4cf197ae'. Despite that generated URL can be set as homepage or used as bookmark, it cannot be yet as search engine in any web browsers, which counts Opera.

Though an URL in GET form, such as 'https://www.startpage.com/do/search?query=%s' , would be a valid one to be used for the addition of a search engine, but it still would use the GET method. In that context, would it be fair to request from Opera the implementation of a new design scheme that would allow a search engine involving such an URL relying on POST method to be set as the default one?