Russian Storytelling Video Dataset for AI Training

This dataset contains high-quality video recordings of 700 native Russian speakers, each speaking freely for 10+ minutes. The content includes full-body video with visible gestures, facial expressions, and natural speech — making it ideal for training multimodal AI models.

Cover

📦 Dataset Overview

700 unique participants
10+ minutes of video per participant
At least Full HD (1920x1080), many samples in 2K and 4K, 30 FPS
Natural speech in Russian (unscripted)
Full-body framing (waist-up or lower)
Visible face, gestures, body language
Clean lighting and audio
Commercial usage rights secured
Includes age and gender metadata

📊 Sample

📺 Preview Video (10 Participants)

To get a quick impression of the dataset, watch this short compilation video (hosted on Google Drive):

🔗 Watch Preview (Google Drive) This sample includes 10 different native Russian speakers, each telling unscripted personal stories with visible gestures, emotions, and clear speech.

See sample: sample_video_dataset_100_participants.csv
Screenshots: 1.png, 2.png, 3.png, 4.png, 5.jpg, cover.jpg
Preview PDF: storytelling_dataset_license_agreement.pdf

🔍 Use Cases

Multimodal model training
Emotion and gesture recognition
Conversational avatars
VR/AR training simulation
NLP and speech-to-text applications

💼 Licensing

This dataset is provided under a non-exclusive commercial license. Redistribution is prohibited.
See: storytelling_dataset_license_agreement.pdf

💬 Contact

Email: [email protected]
Telegram: @Marat_DV

📌 Описание на русском

Набор видеоданных с 700 русскоязычными участниками, рассказывающими истории свободным текстом. Видео в высоком качестве (Full HD и выше), съёмка по колено, видны жесты, мимика, эмоции. Идеально подходит для задач обучения моделей речи, видеоаналитики и распознавания эмоций.

Каждое видео: 10+ минут, 30 FPS, Full HD и выше
Участники дали разрешение на коммерческое и AI-использование.

📩 Контакт: [email protected]
📬 Telegram: @Marat_DV