Call Center Audio Datasets Needed English , Hindi , Spanish

We are currently looking for pre-recorded call center conversation datasets with the following requirements:

• English – 500 hours
• Hindi – 500 hours
• Spanish – 500 hours

Requirements:

* Agent-side audio only
* No background noise
* No long silence segments
* Audio format: WAV
* 16 kHz, 16-bit or higher
* Unidirectional audio
* Transcription text required

Additionally, please confirm the following details:

1. Is transcription text available?

* If yes, is it AI-generated or human-annotated?
* If AI-generated, can it be manually refined?

2. If transcripts are available, please share details regarding:

* Speaker tags
* Timestamp information
* Accuracy rates for:
• Speaker diarization
• WER (Word Error Rate)
• Timestamp alignment

3. Please confirm the audio sampling rate specifications.

4. Has personal information within the recordings been de-identified?

* If yes, please explain the de-identification process.

5. Scope of Data Use:

* The dataset will only be used for internal model training purposes.