Pexo Video Generator for Openclaw

An AI agent skill powered by Pexo that transforms plain-text descriptions into fully compiled, publish-ready videos including scripts, music, and subtitles.

pexo
v0.1.1
Jun 8, 2026
0
583
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install make-a-video

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install make-a-video using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Pexo Video Generator?

The Pexo Video Generator skill empowers AI coding agents to create fully edited videos from simple, natural language prompts. Instead of micro-managing video creation or editing, this skill serves as a relay to the hosted Pexo service. Openclaw Skills leverage this backend to automate high-fidelity asset creation. Pexo's AI handles the entire creative pipeline—including scriptwriting, music selection, voiceover models, subtitle generation, and transitions—resulting in professional video content dynamically.

By serving as an agent-friendly wrapper, it enables developers to prompt video creation within terminal sessions or multi-agent environments. The skill utilizes lightweight bash scripts to communicate with Pexo's robust REST API, managing files, orchestrating conversations, and rendering assets seamlessly without needing heavyweight rendering infrastructure on the host machine.

Pexo Video Generator Use Cases

  • Generating marketing videos and social media promos from raw text ideas
  • Producing video walkthroughs or visual demos from scripts and image assets
  • Quick-turnaround video drafts in automated workflows without manual editing
  • Extending AI agent capacities to produce rich multimedia output using specialized Openclaw Skills

How Pexo Video Generator Works

  1. Project Initiation: The agent runs pexo-project-create.sh with a project brief, initializing a unique session on Pexo's servers.
  2. Asset Ingestion: If local assets (images, audio clips, or reference videos) are provided, the skill uploads them via pexo-upload.sh and generates an asset tag wrapper (e.g., <original-image>asset_id</original-image>).
  3. Request Dispatch: The agent sends the user's creative prompt and asset tags to Pexo's processing pipeline using pexo-chat.sh.
  4. Status Polling & Interaction: The agent continuously queries pexo-project-get.sh to track execution states. During this stage, it handles preview options, answers questions from the Pexo model, or continues the run if there is a disconnect.
  5. Asset Delivery & Refinement: Once processing is complete, the skill fetches the final, high-definition output URL using pexo-asset-get.sh. It also supports feedback iterations (revisions) on the same project using the same session.

Pexo Video Generator Setup

To get started with Openclaw Skills for Pexo, configure your API credentials and environment:

  1. Ensure you have the required runtimes installed: curl, jq, and file.
  2. Obtain an API key from pexo.ai.
  3. Create your configuration directory and file at ~/.pexo/config:
    PEXO_BASE_URL="https://pexo.ai"
    PEXO_API_KEY="sk-<your-api-key>"
    
  4. If this is your first run, execute the diagnostics script to ensure setup and dependencies are correct:
    ./scripts/pexo-doctor.sh
    
  5. If you run into authorization issues, review the instructions in references/SETUP-CHECKLIST.md or check your API key validity.

Pexo Video Generator Data Schema & Taxonomy

This skill leverages a clean schema to keep track of tasks, processing states, and visual assets:

Project State Machine

  • WAIT: The project is rendering or queueing. The agent waits and polls periodically without resending requests.
  • RESPOND: Interaction is needed. The agent handles text answers, picks visual options (A/B/C), or acknowledges documents.
  • DELIVER: Production completed successfully. The agent retrieves the final production asset URL.
  • FAILED: Rendering failed. Diagnostic messages provide a path to retry.
  • RECONNECT: Re-initiates lost connection with a continue flag to resume rendering context.

Script Schema Reference

Script Role Parameters Returns
pexo-project-create.sh Initializes a production session "<brief>" project_id
pexo-upload.sh Uploads reference files to the project <project_id> <path> asset_id
pexo-chat.sh Dispatches prompts & selections <project_id> "<prompt>" [--choice <id>] API Acknowledgement
pexo-project-get.sh Retrieves the latest pipeline state <project_id> Project state JSON
pexo-asset-get.sh Fetches temporary and final output URLs <project_id> <asset_id> Asset URL details
pexo-doctor.sh Runs setup verification and health checks None Diagnostic log

Pexo Video Generator Advanced Features

  • Interactive Human-In-The-Loop Choices: Supports mid-production rendering choices (e.g. choosing between preview A, B, or C) directly through standard API handshakes inside the Openclaw Skills pipeline.
  • Asset Tagging Engine: Auto-references files through strict XML-like tags (e.g., <original-image>) to guide the creative model during layout assembly.
  • Stateful Video Revisions: Supports sequential refinements on the exact same project context (preserving history) rather than resetting the workspace.
  • Fail-safe Reconnection Handling: Detects dropped connections during lengthy renderings and automatically resumes progress via state recovery requests.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*