This dataset consists of Spanish-language telecom call center recordings featuring interactions between Mexican Spanish-speaking customers and agents. The data captures real-world customer service conversations in the telecommunications domain. The dataset is audio-only, without transcripts or annotations, making it suitable for organizations looking to build or customize their own speech and language models from raw conversational data.
| Name | Type | Nullable | Description |
|---|---|---|---|
| uique id | string | No |
The dataset includes a collection of telecom-related customer support calls conducted in Mexican Spanish, reflecting authentic communication patterns found in real call center environments. These conversations may include inquiries related to billing, service activation, troubleshooting, plan upgrades, and general customer support. The audio captures natural conversational dynamics, such as regional accents, speech variations, pauses, interruptions, and call flow structures typical of telecom operations. As an unannotated audio dataset, it provides flexibility for organizations to apply custom labeling, transcription, and annotation pipelines tailored to specific AI/ML use cases.
Real-world call center recordings Natural speech variability (accent, tone, pace) Includes realistic conditions: Background noise Call artifacts Overlapping speech No transcription or annotation included May require preprocessing for: Noise reduction Segmentation Speech-to-text conversion
This dataset contains real-world or simulated US-English (US-US accent) call center conversations from the telecom domain.
License: NA — This dataset is licensed for use by the purchasing entity. Refer to the specific license terms for permitted use.
Access Control: Download links are time-limited and single-use. All access is logged for audit purposes.