GNU ELPA - jit-spell

To enable spell checking in a buffer, type M-x jit-spell-mode RET. To correct a misspelling, you can right-click the word (assuming you have context-menu-mode activated) or call the command jit-spell-correct-word, which uses the minibuffer to read a correction or accept the word.

To make your settings permanent, you may wish to add some variant of the following to your init file:

(add-hook 'text-mode-hook 'jit-spell-mode)
(add-hook 'prog-mode-hook 'jit-spell-mode)
(with-eval-after-load 'jit-spell
  (define-key jit-spell-mode-map (kbd "C-;") 'jit-spell-correct-word))

To pick a spell checker and dictionaries, jit-spell uses Emacs's built-in ispell support code. For instance, ispell-change-dictionary temporarily changes the dictionary of the current buffer. Try also M-x customize-group RET ispell RET to see a listing of all possible settings.

Often there are regions of the buffer that should be ignored for spell-checking purposes. In most cases, no additional configuration is necessary.

The simplest mechanism to make adjustments is the user option jit-spell-ignored-faces. Any word fontified with one of these faces in this list ignored by jit-spell. To find out which faces are present on a given character, you can use the describe-char command.

In all programming language modes, spell checking is restricted to comments, docstrings and strings. This can be modified by customizing the variable jit-spell-prog-mode-faces.

Discussions, suggestions and code contributions are welcome! Since this package is part of GNU ELPA, contributions require a copyright assignment to the FSF.