Question: I need a tool that can perform natural language search in personal videos and images; do you know of any self-hosted web apps that can do this?

Twelve Labs is a sophisticated multimodal AI-powered video understanding platform. It lets you search, generate text and classify large video catalogs with natural language processing. The platform offers APIs for searching specific scenes, generating text summaries and automatically classifying video content. It's designed for scalability, high accuracy and enterprise-level security, so it's a good option for companies with big video archives.

Another alternative is Knowbase, which combines file storage with AI-powered chat. You can store and retrieve files, including videos and audio recordings, by asking conversational questions. It can handle a variety of file formats and offers features like centralized libraries, transcription of video and audio recordings, and sharing. Pricing is based on storage capacity and query limits, so it's good for individuals and teams.

If you're looking for something more video focused, check out Imaginario. This AI-powered video content management platform lets you search multimodal video elements like dialogue, actions, sounds and themes. It also offers auto-framing, AI transcription and collaboration tools so you can manage and share video content without manual tagging or metadata. It offers a free tier and several paid plans.