Audio · Video · Multimodal / AI & Machine Learning
Speech Data | 20,000+ Hours of Real Job Interview Audio & Video | Worldwide | Consented & Unscripted
Advertised by providerLicensing rights not independently verifiedPrice on request
Overview
Speech Data | 20,000+ Hours of Real Job Interview Audio & Video | Worldwide | Consented & Unscripted is a multimodal dataset. The provider states a volume of 20,000 hours. Typical applications include speech recognition, LLM pre-training and conversational AI. Relevant sectors: AI & Machine Learning and Education. Pricing is not published and is provided on request.
Primary applications
What the listing offers
- Stated volume of 20k hours.
What this dataset is the provider licenses real job interview recordings captured on its own AI interviewer platform, used by 4,000+ companies. Every recording is a genuine, high-stakes conversation between a real candidate and an AI interviewer - not scripted prompts, not read speech, and not acted dialogue. The collection currently stands at 20,000+ hours of synchronized audio and video, and it grows by 100+ hours every day. How it is sourced Candidates give explicit consent after their interview, specifically for AI training use. Nothing is scraped and nothing is repurposed from public video. Before you commit to anything, we can share the exact consent language, an explanation of the collection process, the available metadata and dataset structure, and the permitted usage scope. Why interview speech is different Because people are under real pressure and genuinely care about the outcome, these recordings capture natural speech behaviour that studio-recorded and read-speech corpora do not: - Pronunciation and regional accents - Speaking pace and fluency - Pauses, hesitation, and self-correction - Emotional tone and communication style Speakers come from different countries, languages, and accent groups, including communities that are commonly underrepresented in commercial speech datasets, with particular depth in Global South accents. Coverage Recordings are delivered in seven regional segments - Africa, Europe, North America, the Middle East, South America, Southeast Asia and
Related datasets
More in Video & Physical AI →Audio · Text · MultimodalAdvertised
Podcast Audio Data | Egyptian Arabic Multi-Speaker Dialogue Audio Corpus | 10,000+ Hours with Full Transcription | .WAV .JSON
Specifications: ‣ Volume: 100,000+ hours ‣ Format: 8kHz, mono channel (WAV audio + JSON transcript) ‣ Environment: Studio, podcast-quality recording ‣ Recording content:…
- Language
- Arabic
- Volume
- 10k h
- Domain
- Education +1
- Use case
- Speech recognition
Price on request
More speech & audio datasets
All 1,727 speech & audio datasets →- Speech Data | 25,000 Hours of Customer Support Calls with Transcripts for ASR & Voice AI
- Speech Data | OTS 100,000+ Hours of Free Speech Global + NICHE LANGUAGES | GDPR, CCPA Compliant | Native Speakers and Proprietary Data
- Speech Data | OTS 30,000+ Hours of Free Speech Global Accents of ENGLISH | GDPR, CCPA Compliant | Native Speakers and Proprietary Data
- Speech ML Data | On demand Hours of Text-To-Speech (Hard-to-Source Languages) | GDPR, CCPA Compliant | Native Speakers 180+ Countries
- Speech ML Data | OTS 25000+ Hours of Free Speech NICHE LANGUAGES | GDPR, CCPA Compliant | Native Speakers and Proprietary Data
- Speech ML Data | OTS 3000+ Hours of Free Speech URDU | GDPR, CCPA Compliant | Native Speakers and Proprietary Data
- Speech ML Data | OTS 5000+ Hours of Free Speech Swahili | GDPR, CCPA Compliant | Native Speakers and Proprietary Data