Vidrune processes videos entirely in users' browsers using our VISE engine, which analyzes videos frame-by-frame using WebGPU-accelerated AI models.
website
Crystalrohr is building Vidrune, a decentralized video indexing platform that makes video content as searchable as text, without centralized control.
Vidrune processes videos entirely in users' browsers using our VISE engine, which analyzes videos frame-by-frame using WebGPU-accelerated AI models. It costs ~$0.02 per hour of content and runs completely locally
Multiple indexers process the same video independently. We compare their results, reward accurate submissions, and use conviction markets plus AI judges to resolve error disputes. This creates cryptoeconomically guaranteed accuracy that self heals the network.
We intend to make our revenue through: Semantic video search - Find specific moments by what's actually happening in the video, not just titles and tags.
Accessibility - Auto-generated audio descriptions with environmental context, emotions, and gestures for visually impaired users.
Video intelligence API - Real-time access to indexed content for social media monitoring, content moderation, research, and AI training.
we've created our first client called vidrune and will keep integrating more IP related tracking into it
Actively looking to raise.