preedit

Preedit is the provisional text produced by an input method editor (IME) to represent the characters a user is composing but has not yet inserted into the document. It is a central concept in languages that require character-level disambiguation, such as Chinese, Japanese, and Korean, and in other text-entry systems that use composition before finalization. As the user types, the IME updates the preedit string, optionally presenting a list of candidate conversions and allowing the user to select one or continue typing to refine the composition.

The preedit is typically rendered differently from committed text. Common approaches include underlining, coloring, or highlighting portions of the preedit to indicate current segments, tones, or cursor position. The application receives the final text only when the user confirms the choice, at which point the preedit is committed and replaces the provisional string.

With many input methods, the user can insert or adjust punctuation within the preedit, switch between candidate forms, or move the cursor within the composition. The concept also encompasses the internal state of the input framework, where a preedit string, its attributes, and its position are maintained separately from the document model.

Examples include pinyin input for Chinese characters, kana-to-kanji conversion in Japanese, and Hangul composition in Korean. Preedit is supported by input method frameworks and operating system IMEs, which provide APIs to set the preedit text and its styling for rendering in applications.