AI Video Creation Workspace
LibTV is a video creation platform built by LiblibAI, serving two types of users: creators working in a browser and AI agents that call it through APIs.
On the creator side is a visual infinite canvas, where scripts, characters, storyboards, images, videos, and audio assets are all organized within the same project, and you can open branches to test different visual directions. On the model side, it integrates Seedance 2.5, Wan 3.0, MiniMax H3, and others, plus proprietary features such as Director’s Desk, which provides a 3D virtual scene for precise control over spatial relationships; frame-by-frame reference, which lets you upload a video and use it as frame-by-frame reference; and segment retake, which allows you to modify only a specific segment of a video.
The developer side provides an official CLI, Skills, and OpenAPI. Agents can create creation sessions, send image and video generation instructions, upload assets, check task progress, and batch download results—ideal for studios and content teams looking to build automated video pipelines. Regular web users don’t need to deal with this part.
Billing is credit-based. Generation costs are calculated based on model, duration, and specifications, and members receive discounts. Promotional pricing changes frequently, so before purchasing, we recommend checking the current day's credit usage rules and validity period on the checkout page.
Before shooting a vertical short drama, you can check TVLab to see how hit dramas in the same genre open and which episode the twist happens in, then pick an inspiration card with pre-split storyboards and generate shot by shot.