A high-fidelity video production pipeline that generates 30-second cinematic ads with consistent character appearances across multiple scenes.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install xpilot-ad-maker
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install xpilot-ad-maker 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 xpilot-ad-maker?
The xpilot-ad-maker is a sophisticated automation tool within the Openclaw Skills ecosystem designed for professional-grade video marketing. It solves the common 'character drift' problem in AI video generation by leveraging Vidu reference-to-video technology, ensuring a single protagonist remains visually consistent across four distinct storyboarded shots.
This skill orchestrates a complex multi-model workflow involving Wavespeed for character seeding, Vidu for cinematic motion, and Replicate Kokoro for high-quality TTS narration. By automating the post-production process—including ffmpeg-based brand overlays, ambient music mixing, and Cloudflare R2 storage—it reduces the cost and time of creating high-impact video advertisements to roughly $2.55 and 3 minutes per run.
xpilot-ad-maker Use Cases
- Creating localized medical tourism advertisements with high-quality visual storytelling.
- Rapid prototyping for social media video campaigns requiring character continuity.
- Automated production of multi-scene brand stories for digital marketing agencies.
- Developing consistent video content for verticals like dental tourism, cosmetic surgery, or travel agencies.
How xpilot-ad-maker Works
- Character Seeding: Generates a high-fidelity protagonist portrait using Wavespeed Seedream 4.5 and mirrors it to R2 storage.
- Parallel Video Generation: Submits four parallel Vidu reference2video requests to create distinct scenes featuring the seeded protagonist.
- Narrative Synthesis: Uses Replicate Kokoro TTS to generate time-aligned narration for each specific shot.
- Video Assembly: Concatenates the four clips into a single 30-second silent master file using ffmpeg.
- Post-Production Overlay: Applies complex ffmpeg filters to add brand captions, top descriptive text, and a synthesized ambient audio track.
- Final Distribution: Uploads the finished MP4 and all intermediate assets to a Cloudflare R2 bucket for immediate deployment.
xpilot-ad-maker Setup
Ensure you have node (>= 18) installed. Configure your environment variables for VIDU_API_KEY, WAVESPEED_API_KEY, REPLICATE_API_KEY, and your Cloudflare R2 credentials.
# Clone the Openclaw Skills repository and navigate to the skill
cd skills/xpilot-ad-maker
# Install dependencies
npm install
# Execute the full ad generation pipeline
npx tsx make-xpilot-ad.ts
xpilot-ad-maker Data Schema & Taxonomy
The skill manages a structured asset pipeline through Cloudflare R2 and local temporary storage:
| Asset Type | Source/Model | Purpose |
|---|---|---|
| Reference Image | Wavespeed 4.5 | Establishes character identity |
| Shot Clips | Vidu q2-pro | 4x parallel 7-8 second video segments |
| Audio Clips | Replicate Kokoro | Shot-specific .wav narration files |
| Final Master | ffmpeg | 30s MP4 with music and text overlays |
| Metadata | R2 Public URL | Traceable links for all intermediate assets |
xpilot-ad-maker Advanced Features
- Post-Production Iteration: Run xpilot-ad-finalize.ts to update captions or audio without re-spending Vidu credits.
- Vertical Customization: Easily swap prompts, narration, and brand colors to pivot from medical tourism to any other industry.
- Reference-to-Video Mode: Built-in logic to handle complex Vidu URL signatures and signature persistence.
- Automated Rate Limiting: Intelligent 11-second delays for TTS calls to stay within Replicate free-tier or low-credit quotas.
- Asset Traceability: Every intermediate clip is mirrored to R2 for manual re-editing if needed.
SKILL.md
Loading
Related Openclaw Skills
Weather Fetcher Skill
A Python-based skill for retrieving detailed real-time weather reports, including AQI and lifestyle advice, for cities across China.
Seedance 2.0 Prompt Engineer
An expert AI skill designed to convert natural language descriptions into high-fidelity, structured prompts for the Seedance 2.0 video generation platform.
US Treasury Radar
A sophisticated monitoring tool that tracks US Treasury debt supply and demand to identify macroeconomic risk signals and liquidity trends.
GitHub
A powerful skill enabling AI agents to manage GitHub repositories, pull requests, and CI/CD pipelines using the official GitHub CLI.
Leaptic
Leaptic provides an Openclaw Skills integration to access real-time device snapshots, battery status, and media storage data for your hardware.
Archive Intelligence
An automated intelligence gathering skill designed to monitor, collect, and report on the archives and records management industry.
Featured*
