< (Shift + ,): Previous tree (saves current edits)
> (Shift + .): Next tree (saves current edits)
Shift + S: Export all edited trees as .conllu
Selecting tokens
Click any word or POS cell to select it.
Selected token is highlighted; in Mass Edit mode, the anchor (head word) stays highlighted.
Single-word edits (choose a word, press a key)
r: Make the selected word w1 the root of the entire sentence. Add a DEPREL where w1 is the head and the old root the dependent.
v / n / a / i: Set POS = VERB / NOUN / ADV / INTJ
Two-word relations: click on w1/head, press a key, then click on w2/dependent
Press b: Set w2 as w1’s subject (nsubj; csubj if w2 is a VERB). Any existing subject of w1 becomes fmrsubj.
Press c: Set w2 as w1’s complement:
dobj if w2 is a NOUN
xcomp if w2 is a non‑finite VERB (VerbForm=Inf or “to”)
CCOMP if w2 is a finite VERB
Press p: w2 is in a discourse deprel with w1 (e.g. sentence-final particle)
Press t: w2 is in a dislocated deprel with w1 (e.g. topics)
Press x: w2 is in a parataxis deprel with w1 (e.g. w2 is actually the main verb of a second sentence)
Mass Edit mode: Suppose we have a word w1. Mass Edit lets us identify the words that take w1 as their head and the corresponding DEPRELs.
Press m: Enter Mass Edit (status shows “active”).
Click w1 - this is now the head word. Press one of b/c/p/t/x (the DEPREL), then click w2 to apply. Now w2 is in this DEPREL relation with w1. Repeat with new letters to link w1 to w3, w4, etc.
Esc (once): Clear pending actions but stay in Mass Edit. Click a new w1 to continue.
Esc (again while idle): Exit Mass Edit.
Split / Merge (token boundary edits)
- (hyphen): Split selected token. A one-line input appears; type space-separated words (at least two words). Each becomes a new token with POS=VERB.
+ (plus): Merge with an adjacent token (click it). Forms are concatenated (no space), POS=UNK. Non‑adjacent clicks are rejected.