Commercial training data from specialist providers. Listings show what each source advertises; licensing is confirmed with the provider before any agreement.
Scope: Synthetic customer–agent interactions, available in dual-channel (per speaker) or single-channel mixes.
Scope: Synthetic customer–agent interactions, available in dual-channel (per speaker) or single-channel mixes.
Scope Short, prompted utterances and wake words such as “Go,” “Stop,” “Hey Siri”, and domain-specific command phrases.
Two-speaker spontaneous conversations on everyday topics, collected on demand from our 1.2M+ global users across 180 countries.
Clean, high-quality scripted recordings aligned with transcripts, built for training expressive and natural-sounding TTS systems.
Scope: Short commands and wake words (“Go”, “Stop”, “Hey Siri”), with each user providing 150–200 utterances.
Two-speaker scripted conversations on everyday topics, collected on demand from our 1.2M+ global users across 180 countries.
Scope Short, prompted utterances and wake words such as “Go,” “Stop,” “Hey Siri”, and domain-specific command phrases.
Clean, high-quality scripted recordings aligned with transcripts, built for training expressive and natural-sounding TTS systems.
Two-speaker spontaneous conversations on everyday topics, collected on demand from our 1.2M+ global users across 180 countries.
A comprehensive collection of 45,000+ authentic 4K sports video clips captured across real-world shoots, multiple sporting disciplines, environments and participant…
This world model dataset is produced on demand based on the buyer’s required tasks, sensor stack, participant profiles, capture environments, annotation depth and target…
Knives are small, thin and often half-hidden in a hand, which makes them one of the hardest objects for security cameras to catch.
Fire detection has to work in the first minutes, when the smoke is thin and the flames are small, and it must not alarm on every steam vent or floodlight.
Weapon detection models fail in the real world for two reasons: too few training images of guns seen from real CCTV angles, and too many false alarms on everyday objects.
For decades, a game's manual was part of the game.
1. Overview Date size: 1 million sets Image resolution: matches device resolution App types: built-in apps, shopping, food ordering, education, social, video, games, etc.
"Collection of 100,000 high-quality video clips across diverse real-world domains, designed to accelerate the training and optimization of computer vision, machine…
The provider offers a broad collection of high-quality annotated imagery data and AI-training video datasets, comprising 15,000+ 4K/6K RAW assets captured on RED cinema…
3D CAD Visual Rendering Dataset – The 3D CAD Visual Rendering Dataset is a professionally curated collection of high-quality three-dimensional computer-aided design…
The provider delivers diverse computer vision and AI training datasets with over 4,000 assets in 4K/6K RAW video.
Current standard package highlights: - Expert trajectories across NVIDIA classic indoor scenes: Hospital, Office, Warehouse, Simple Room - Diverse route semantics:…
The provider delivers diverse computer vision and AI training datasets with over 20,000 assets in 4K/6K RAW video.
METADATA COLLECTED: Session (automatic): session_name, start_time_unix, episode_duration_s, fps, firmware_version, encryption Per-clip capture log: operator_id, task_id,…