Pexo launches ClawHub video skill that auto-picks Sora or Kling and returns edits in chat
Pexo went live on ClawHub as an OpenClaw skill that builds complete videos inside chat, asking clarifying questions and auto-selecting models scene by scene. It matters if you want ads or explainers without opening a separate editor, but review the storyboard before posting.
TL;DR
- Pexo is now live as the first full video-creation skill on ClawHub, so an OpenClaw agent can return a finished video inside chat instead of sending you to a separate editor, according to the launch thread.
- The distinctive part is model routing: the demo thread says Pexo silently chooses scene-by-scene between tools like Sora, Kling, and Seedance rather than making the user pick a generator.
- In one creator test, a rough one-message product-ad brief triggered two clarifying questions, then a storyboard, rendered scenes, and music without any manual editing.
- Pexo is pitched as both a ClawHub install and a standalone product; the install note says the OpenClaw version adds video generation with no new UI, while Pexo's site describes ready-to-post outputs with voiceover, captions, transitions, and music.
What shipped
Pexo launched on ClawHub as an OpenClaw skill that turns a chat agent into a video-production tool. The core promise in the main announcement is workflow compression: describe the video in Telegram, Discord, or WhatsApp and get back a completed edit rather than a short raw clip.
That matters because Pexo is not framed as a single-model wrapper. In the model-selection post, the product is described as automatically choosing the best model for each scene, including Sora, Kling, or Seedance, with that routing hidden from the user. Another thread post adds that the returned output already includes transitions, background music, and pacing, positioned as a ready-to-post asset rather than footage for a second editing pass.
How the chat workflow works
The workflow Pexo is selling is conversational production, not prompt engineering. The feature overview says the user describes what they want in plain language and the agent handles the build, while a hands-on example claims a rough product-ad request led to two clarifying questions before Pexo generated a storyboard, rendered scenes, and added music.
The standalone product page at Pexo's site expands that pitch: users can upload assets, iterate through feedback, and get a full video with layers such as captions and voiceover. The OpenClaw version keeps that process inside an existing agent workflow instead of moving creators into a separate timeline editor.
Where it fits for creators
The strongest creative use cases in the launch material are lightweight commercial formats. The use-case list calls out product ads, UGC-style content, social clips, and explainer videos, which all benefit from fast scripting, scene assembly, and music selection more than frame-precise manual editing.
The practical appeal is less “best possible cinematic control” than fewer tool hops. The workflow post frames Pexo as video creation folded into the same agent already handling other tasks, and the install note says setup is a one-step ClawHub install with no new subscriptions or interface to learn.