d.velop documents vs Transkribus

d.velop documents has no write-up on Specifiction yet; its figures are above.

Transkribus uses AI to convert handwritten and printed documents into searchable text. Its recognition supports more than 100 languages and historical scripts. Users can select public models for handwriting, print, layout analysis and table recognition, or train their own. The editor supports text and layout changes, version history and real-time collaboration. Layout analysis identifies text regions, baselines, tables and reading order, with exports including PAGE XML, ALTO and searchable PDF. The Developer Platform API returns structured results as JSON, a ZIP archive or PAGE XML; it is in beta and requires Scholar or a higher plan. Results remain available for 24 hours, and submitted JPEG, PNG or TIFF images may be up to 20 MB. Document processing consumes credits: several recognition tasks cost 1 credit per page, while line detection costs 0.25 credits per page. Free includes 50 credits per month; Scholar costs €99 a year. Transkribus says documents are processed and stored on its servers in Austria, and users own and can delete their documents and transcriptions.