Skills

Use pexo-agent to turn text or assets into a finished video

For people who want an agent to do AI video generation. Install pexo-agent first. Inputs can be text, images, URLs, scripts, or audio; the output is a finished video. You need only one API key.

Official source · pexo-skills

Install

npx skills add https://github.com/pexoai/pexo-skills --skill pexo-agent

Who it's for

Use this skill when you want an agent compatible with Agent Skills to do AI video generation. The source writes it as an AI video generation skill: it automatically picks among more than ten models; inputs can be text, images, URLs, scripts, or audio; the output is a finished video. Official repository:

https://github.com/pexoai/pexo-skills

Images and audio are inputs here, not another class of output beside video. Do not read the repository title as a skill set that makes images, sound, and films in parallel.

Preconditions: the current agent can receive Agent Skills, and you accept that a real generation needs one API key. If you are not ready to run the official install on this machine, or you are not ready to provide that key, do not continue yet. The source writes that prompt engineering is not needed, and video editing is not needed. If you only want a still image or audio alone and do not want a finished video, this is also not where to start.

The source named several clients and also writes that any agent that supports the Agent Skills standard can use it. Do not copy that roster into the body as an exclusive scope. Standard notes:

https://agentskills.io/

First try in 3 minutes

First confirm the current agent can receive Agent Skills. Open the official repository and check:

https://github.com/pexoai/pexo-skills

Recommended install is only this one. Installing the skill writes files on this machine; before you run it, confirm you want pexo-agent left here:

npx skills add https://github.com/pexoai/pexo-skills --skill pexo-agent

The address in the command must match the official repository exactly. Success signal: this command ends normally. The source does not write a skill landing directory; do not invent a path. The source also has another install command aimed at a particular client; the first time, do not put it beside this as a second main path.

The first time in the agent, only check; do not generate a video yet. Ask the agent to state: this is AI video generation; inputs can be text, images, URLs, scripts, or audio; the output is a finished video; only one API key is needed; prompt engineering is not needed; video editing is not needed. The observation is that it can recite inputs in and finished video out, has not gone online, and has not started model picking and generation.

A real generation calls external models and writes a video; that is a side effect. Continue only when you confirm you will use your own API key and accept getting a finished video. The source does not write a key field name; do not write a placeholder key. The first time, do not invent editing steps, and do not invent a model roster.

Use cases

Case 1: First install pexo-agent via the official path. In an agent compatible with Agent Skills, after you accept writing skill files, only run npx skills add, address

https://github.com/pexoai/pexo-skills

with --skill pexo-agent. Fits people who only want the video generation skill attached to an agent and are not ready to produce a film. Success signal: the install command ends normally, and the agent can say finished video comes out. If install fails, stop; do not switch to the install command that was not chosen as the main path.

Case 2: First check inputs, outputs, and the key, then decide whether to generate. After install, have the agent recite: text, images, URLs, scripts, or audio can all be inputs; the result is a finished video; only one API key is needed; it automatically picks among more than ten models; prompt engineering is not needed; video editing is not needed. Fits people who worry the agent will first make images, make sound, or start editing. Success signal: it states inputs and output correctly. Before a real generation, confirm you will use that API key and accept that this run will write a video. The source gives no prompt template or editing switch; do not invent them.

Boundaries

The main path is only npx skills add; the target skill is pexo-agent; the repository address must be

https://github.com/pexoai/pexo-skills

The source also wrote another install command; the first time, do not put them side by side. Compatibility is written as agents that support the Agent Skills standard:

https://agentskills.io/

Do not copy product names that appeared as the only usable scope. Capability stops at AI video generation: automatic model picking, multiple inputs, finished video out. Images and audio are not parallel outputs. Only one API key is needed; prompt engineering is not needed; video editing is not needed. The source does not write key fields, a model inventory, or uninstall steps. Do not fill in unwritten side effects and operations. First-time success is only install finishing and the agent reciting inputs and output; before a real film you must confirm use of that API key. If the agent starts making images, making sound, or opening an editing timeline, stop first and return to finished video as the one output. The source gives no copyable generation prompt; the first time, do not invent storyboard or editing parameters. More than ten models are written only as automatic picking; do not write unwritten model names as options.