What LibTV is
LibTV is an AI video creation tool for video creators, built by Evoken (演语科技), the AI application company founded by Chen Mian. Its official site is liblib.tv, titled “专业视频创作工具” (a professional video creation tool). Per the official site, it combines a canvas-style creation workflow with generation and editing: entries for video, image, and audio generation plus smart editing, alongside an assistant called LibTV Agent. The site also lists specialized features such as deep motion capture, frame-by-frame review, a director’s console, clip re-shooting, and talking-head videos, plus extensions like a Blender plugin.
LibTV, Liblib (the AI image-creation and model-sharing community), and Lovart (the AI design agent) are three separate product lines of the same company — and the domains differ: LibTV uses liblib.tv while Liblib uses liblib.art.
Release timeline
Per Chen Mian’s own account in the LateTalk episode 175 interview (July 2026, republished by Chwang), LibTV launched in March 2026, and he says it contributes more than half of company revenue; that revenue figure is a founder’s statement. Current features and plans are on the official site.
Using it and its limits
The official site’s entry points are new projects and new canvas creations; generation and editing combine inside the canvas workflow. The underlying models available for video generation, the feature set, and usage rules change with versions, so the site and in-product documentation are the reference before use.
Discussion in the show
Neither published Weekly transcript names LibTV directly, but two passages touch on its field:
- In Weekly #001’s chapter “AI 应用市场回暖了吗？利润空间比需求更现实” (is the AI application market recovering?), Orange cites “like Chen Mian, who packages it into some plan” as an example of bundling model costs into subscription plans and betting on usage (the line in the Chinese transcript). Chen Mian is the founder of the company behind LibTV; “Mian Shen” is the hosts’ nickname for him (see also the token-selling chapter in the same episode).
- In Weekly #002’s chapter “音乐与图像编辑：可控修改的进步与局限” (music and image editing), Yang Pan says “didn’t we get agent-based editing for video and images a year ago?” (the line in the Chinese transcript), comparing how quickly different content industries adopted agent-based editing — a judgment about the field as a whole, with no product named.
These are the speakers’ judgments in the show’s context, not a review of LibTV.
Frequently asked questions
What is LibTV?
An AI video creation tool from Evoken that combines a canvas workflow with video, image, and audio generation, smart editing, and the LibTV Agent assistant. See the official site.
Where is LibTV’s official website?
The official site is liblib.tv. Note this is a different domain from the image community Liblib at liblib.art.
Which company makes LibTV?
Evoken (演语科技), founded by Chen Mian; the same company also makes Liblib and Lovart.
How are LibTV and Liblib related?
They are separate product lines of one company: Liblib (liblib.art) is an AI image-creation and model-sharing community, while LibTV (liblib.tv) is an AI video creation tool. The similar domains cause constant confusion in search.
When did LibTV launch?
Per Chen Mian’s July 2026 interview, LibTV launched in March 2026.