samsar

Samsar is a generative video and media automation platform for turning prompts, product images, structured content, and creative ideas into finished videos and reusable studio assets. It integrates one-shot text-to-video up to 3 minutes, image-list-to-video, image editing, audio generation, semantic search, recommendations, assistant workflows, and multi-provider model routing across a Studio and API surface. The platform supports various AI models across modalities, including inference and vision, image generation, video generation, speech, music, lip sync, sound effects, and embeddings. It allows multi-provider routing so each operation can use the best available provider and model. Users can interact with Samsar through a hosted Studio application, a JavaScript client (samsar-js), or a standalone local deployment using Docker. The standalone deployment includes a setup wizard that configures providers, services, data storage, domains, and admin settings. The system features detailed post-production controls, one-click scene re-rolls after render, sandboxed local logging, and built-in search for creating a generative video library.