Leaptest focus seems to be as a visual recognition type tools.
These tools are usually fairly limited in functionality and create tests that are fragile and easily broken by page changes.
When using selectors to identify parts of the web page, currently a human programmer will make the best decision, ensuring that the locator is robust, unique and not tied into the page layout
Example: Sometimes table tr td td td
would be auto generated by such a tool as a unique identifier but which table cell to use will actually change over time as the table output, layout and content gets changed and what would work much better would be table#todays_selection td.edit_current_amount
, i.e. using classes and ID's to narrow down the right element. This assumes that suitable classes, ID's, data attributes, etc. exist on the page. If they don't exist or are used sparingly you will have to be selective in which you use, will require developers to improve identification of elements or, worst choice, you will have to use include layout tags in your tests making them more brittle and likely to break when the pages get changed for non-related reasons.
My advice would be to start using the Firefox seleniumIDE plugin.
When you outgrow it, use ruby-rspec-selenium which is my tool of choice.