DacteloBy the Dactelo team at Taktek.
An Arabic handwriting data inventory with separate licence pools
By 22 October 2026, Dactelo had built a dataset catalogue and conversion scripts for public Arabic handwriting material. The inventory records source links, dataset labels, available text and licence categories. Historical manuscripts, isolated words and modern handwriting remain distinguishable in the records.
That distinction matters for a project starting with Lebanese legal handwriting. A historical manuscript can provide useful letter forms without representing a modern court document. A word dataset can support an experiment without testing full-line reading. Dataset size alone does not answer either question.
Licence handling is equally specific. Muharaf’s public card states a noncommercial share-alike licence; the Egyptian word dataset states CC BY 4.0. Research-only material stays separate from candidates for a commercial model. Catalogue entries also preserve unresolved source-image questions. This work prepared an auditable inventory, rather than clearing every downstream use.