Pashto AI is a community-driven initiative dedicated to building high-quality, human-verified language datasets for Pashto, a language spoken by more than 50 million people yet still critically underrepresented in artificial intelligence.

Through our data platform, data.pashtoai.org, a growing community of more than 100 native-speaker contributors transcribes, segments, reviews, and validates Pashto speech and text. Our work includes drama audio with speaker-labeled, time-aligned transcripts; post-edited voice recordings; a dictionary containing more than 50,000 entries with phonetic information; Pashto–English translation pairs; OCR corpora; and large collections of carefully validated sentences.

Every dataset passes structured quality-control processes, with transparent contributor attribution and fair compensation. As our community, partnerships, and technical capabilities continue to grow, we plan to contribute significantly more Pashto language data and expand into new AI domains in the future.

Our long-term mission is to ensure that Pashto is represented across every area of artificial intelligence,openly wherever possible and through fair, sustainable partnerships with the communities that create and validate the data.