Urdu Dialogue Transcription Project

I have a collection of two-speaker, natural Pakistani Urdu conversations that must be transcribed and aligned with time stamps so the text can be fed into an AI speech-recognition training pipeline.

Your task is straightforward: listen to each recording, capture every spoken word in standard Urdu script, and insert the required time code markers exactly as outlined in the style guide I will share before we start. Because the material will train language models, consistency and accuracy are critical—spoken nuances, hesitations, and casual phrasing all need to be preserved.

All work will be done remotely. I’ll provide:
• Download links to the audio files (clean studio captures, no background noise)
• A simple template showing the time-stamp format and speaker labels
• A short trial clip so you can confirm you’re comfortable with the workflow before moving on to full batches

Deliver back the completed transcripts in UTF-8 text or Word—whatever you prefer—as long as the time stamps remain intact. When each batch is approved, I’ll release the milestone and send the next set of files.

Experience with tools such as Express Scribe, oTranscribe, or any foot-pedal software is welcome but not mandatory; accuracy, reliability, and native-level Urdu comprehension matter most. If this sounds like a fit, let’s begin with the trial clip right away.