For plagiarism, to my mind it is important to separate student work from publication/communication.

In the case of student work, many tools on line are powerful enough to validate that student did not cheated and to evaluate in a minimum of time how much they are able to reformulate a scientific question or idea.

http://www.grammarly.com/

http://www.paperrater.com/

https://www.plagtracker.com

For scientific publication it is more difficult, for example the introduction, how unique should be the description of Alzheimer's disease (96 000 hits in pubmed) or Parkinson's disease (50 000 hits)? How much a western blot description in the material and method should be original (220 000 hits in pubmed)? I guess then apply to plagiarism not really the form but more the content. But are reviewers not their for that, if they are expert in the field of the publication they should be able to check in a reasonable amount of time the originality of the work using literature data base (such as pubmed for example) and not on plagiarism online tool.

For grammar, words or libreoffice include relatively good tools. Proof reading by colleagues, co-authors or native speakers also helps a lot. Else specialize institute or companies are often not that costly (better to stay local in order to have the chance to interact with them).

Best

use viper Plagiarism scanner, its free and easy to use.