Obviously it would be best for Mendeley to just allow to add the langid field for convenient language switching.
If you cannot have that, you might organise your bibliography in .bib files by language.
You then can add the langid field for each entry like this
\DeclareSourcemap{
\maps[datatype=bibtex]{
\map{
\perdatasource{\jobname-en.bib}
\step[fieldset=langid, fieldvalue={english}]
}
\map{
\perdatasource{\jobname-de.bib}
\step[fieldset=langid, fieldvalue={ngerman}]
}
\map{
\perdatasource{\jobname-fr.bib}
\step[fieldset=langid, fieldvalue={french}]
}
}
}
Where we use \perdatasource to restrict the mapping to the appropriate .bib file.
MWE
\documentclass[ngerman,french,english,british]{article}
\usepackage[T1]{fontenc}
\usepackage[utf8]{inputenc}
\usepackage{babel}
\usepackage[style=authoryear, backend=biber, abbreviate=false, autolang=other]{biblatex}
\usepackage{filecontents}
\begin{filecontents*}{\jobname-en.bib}
@book{abook,
author = {Ann Uthor},
editor = {Emily Ditor},
translator = {Tracy R. Anslator},
title = {A Fancy Title},
date = {2015},
}
\end{filecontents*}
\begin{filecontents*}{\jobname-de.bib}
@book{einbuch,
author = {Andrea Torin},
editor = {Herbert R. Ausgeber},
title = {Ein Titel},
date = {2014},
}
\end{filecontents*}
\begin{filecontents*}{\jobname-fr.bib}
@book{livre,
author = {Amelie Uteur},
editor = {R. Édactrice},
title = {Titre},
date = {2014},
}
\end{filecontents*}
\addbibresource{\jobname-en.bib}
\addbibresource{\jobname-de.bib}
\addbibresource{\jobname-fr.bib}
\DeclareSourcemap{
\maps[datatype=bibtex]{
\map{
\perdatasource{\jobname-en.bib}
\step[fieldset=langid, fieldvalue={english}]
}
\map{
\perdatasource{\jobname-de.bib}
\step[fieldset=langid, fieldvalue={ngerman}]
}
\map{
\perdatasource{\jobname-fr.bib}
\step[fieldset=langid, fieldvalue={french}]
}
}
}
\nocite{*}
\begin{document}
\printbibliography
\end{document}

Note that the language is different for all the entries.
Alternatively - thanks to @jon for the idea - you could also add a unique string to a field you can export to. In our example we add [[<langid>]] to the title field (where <langid> should expand to a language identifier biblatex or rather babel/polyglossia can cope with). The language string is then added to the langid field and deleted from the title.
Here, the mapping looks like this
\DeclareSourcemap{
\maps[datatype=bibtex]{
\map{
\step[fieldsource=title, match=\regexp{\s*\[\[(.+)\]\]}, final]
\step[fieldset=langid, fieldvalue=\regexp{$1}]
\step[fieldsource=title, match=\regexp{\s*\[\[(.+)\]\]}, replace={}]
}
}
}
And you add the language to the .bib entry like this
@book{einbuch,
author = {Andrea Torin},
editor = {Herbert R. Ausgeber},
title = {Ein Titel[[ngerman]]},
date = {2014},
}
MWE
\documentclass[ngerman,french,english,british]{article}
\usepackage[T1]{fontenc}
\usepackage[utf8]{inputenc}
\usepackage{babel}
\usepackage[style=authoryear, backend=biber, abbreviate=false, autolang=other]{biblatex}
\usepackage{filecontents}
\begin{filecontents*}{\jobname.bib}
@book{abook,
author = {Ann Uthor},
editor = {Emily Ditor},
translator = {Tracy R. Anslator},
title = {A Fancy Title [[english]]},
date = {2015},
}
@book{einbuch,
author = {Andrea Torin},
editor = {Herbert R. Ausgeber},
title = {Ein Titel[[ngerman]]},
date = {2014},
}
@book{livre,
author = {Amelie Uteur},
editor = {R. Édactrice},
title = {Titre [[french]]},
date = {2014},
}
\end{filecontents*}
\addbibresource{\jobname.bib}
\DeclareSourcemap{
\maps[datatype=bibtex]{
\map{
\step[fieldsource=title, match=\regexp{\s*\[\[(.+)\]\]}, final]
\step[fieldset=langid, fieldvalue=\regexp{$1}]
\step[fieldsource=title, match=\regexp{\s*\[\[(.+)\]\]}, replace={}]
}
}
}
\nocite{*}
\begin{document}
\printbibliography
\end{document}
\DeclareSourcemapto remove it and then add language fields.... – jon Feb 10 '15 at 15:32