A significant portion of African Literature and expressive cultures – from the late 19th century to the present – has been produced as print, audio, video, and (more recently) digital ephemera. From market pamphlets to YouTube videos, the materiality of the African expressive archive is often anathema to the standards and cataloguing practices of formal libraries and collections. This poses challenges to attempts to make these materials visible and findable in a dataverse dominated by literary data on commercial, English-language, novels. This paper addresses the problem of messy and informal data in the work of the African Literary Metadata (ALMEDA) research project, raising questions about the unruly archive and what is at stake when we try to curate it.