Real-World AI Training Video Collection

I’m building a dataset of everyday video footage that will be used exclusively to train computer-vision models. I need contributors who can record authentic, real-world scenarios—anything from routine street scenes to casual moments at home or in the office—captured in a natural, unscripted way.

Here’s what I’m after:
• Short clips (15 – 90 seconds each) shot in landscape at 1080p or higher.
• Natural lighting and ambient sounds preferred; please avoid heavy post-processing or filters.
• No visible trademarks, sensitive documents, or faces without signed consent.
• A brief text file for each clip noting date, location type (e.g., “suburban street,” “kitchen”), camera model, and any noteworthy context (weather, activity, etc.).

I’ll review submissions for diversity of settings, clarity, and compliance with privacy guidelines before accepting the final batch. Once approved, you’ll simply upload the raw footage plus metadata to a shared drive. Feel free to ask if you’re unsure whether a scene is suitable—variety is useful, and I’m open to creative suggestions as long as they reflect genuine real-world environments.