Ready-made Urdu speech datasets

Availability is confirmed per enquiry. Some material can only be licensed within the consent scope the speakers originally agreed to, and we will not stretch that to close a sale.

• Consent scope: does it permit commercial model training, and by a third party?
• Dialect distribution: a corpus that is 90% one urban variety will not generalise across Uttar Pradesh, Telangana, Bihar
• Recording conditions: studio-only data underperforms badly on noisy deployments
• Transcription convention: Right-to-left Nastaliq tooling errors and diacritic loss will otherwise surface as tokenisation noise
• Overlap: check whether the corpus already sits in the public sets you trained on

Indian Urdu specifically, and Dakhini in particular, are absent from public data dominated by Pakistani Urdu broadcast speech.

This is usually the reason buyers move from licensing to commissioning: the free and cheap material covers the easy half of the language.

• Dialect spread across Dakhini (Hyderabad), Lucknawi, Dehlvi, Bihari Urdu rather than a single prestige variety
• The phonetic contrasts that Urdu models actually get wrong: Shares most phonology with Hindi but adds Perso-Arabic phonemes (/q/, /x/, /ɣ/, /z/, /f/) that many speakers merge
• Code-mixed speech transcribed rather than discarded — Spoken Urdu and spoken Hindi are largely mutually intelligible; the distinction is mainly lexical and orthographic. Decide up front whether transcription is in Nastaliq, Devanagari, or both.
• Recording geography covering Hyderabad, Lucknow, Delhi, Bhopal, since dialect follows location
• Utterance types beyond read prompts: Dakhini spontaneous conversation from Hyderabad, Dual-script transcription sets (Nastaliq plus Devanagari), Formal and colloquial register pairs

Fix the script decision before fielding; retro-transcribing a Nastaliq dataset into Devanagari after delivery costs as much as the original transcription pass.