Premium Studio-Grade, Multi-Speaker English Conversational Audio Dataset with Stems + Transcripts - 35K Hours, 2+ Speakers, 500+ New Hours/Month

This premium studio-grade conversational audio dataset delivers over 34.5K hours of multi-speaker English content captured from professional sports and talk studio environments. Each file includes clean mixed-down audio plus individual isolated stems, diarized transcripts at word and utterance level, unique speaker IDs, and sentiment annotations - enabling precise speech modeling, diarization training, and emotion-aware LLMs. The dataset spans thousands of hours of dynamic, domain-specific conversations with 2+ speakers per episode, covering teams, markets, and sports genres across multiple US markets, capturing regional parts of speech, slang, and dialects/accents from over 150+ unique voices. Data is fully rights-cleared and indemnified for AI training, ensuring compliance and commercial safety. New content is continuously added at an average rate of 500+ hours per month, keeping the dataset current and diverse for model refresh and fine-tuning use-cases. Ideal for LLM developers and AI teams training speech-to-text, speaker embedding, sentiment, and multimodal retrieval models. Delivered securely via Google Drive or AWS S3 with metadata in MP3, WAV, JSON, SRT, VTT and TXT formats. Enterprise licensing only - contact the provider for commercial terms or large sample access.