Turn prompts and existing media into production-ready AI videos through a streamlined command-line workflow.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install video-generation-and-editing
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 video-generation-and-editing 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 Video Generation and Editing?
Video Generation and Editing is an AI video production skill for creating, editing, extending, and transforming videos from text, images, reference videos, audio, or first-and-last frames. It supports text-to-video, image-to-video, reference-to-video, video editing, and video extension workflows through the AI Hive OpenAPI platform.
As part of Openclaw Skills, it removes the need to manually implement API requests, media uploads, pricing checks, task polling, and result downloads. Creators, post-production teams, advertisers, and e-commerce operators can generate product videos, TVCs, social commerce content, UGC-style ads, short dramas, comic dramas, and other AIGC video assets from a consistent Python CLI.
Video Generation and Editing Use Cases
- Generate cinematic videos from natural-language prompts with text-to-video.
- Animate product images, storyboards, or key visuals with image-to-video.
- Maintain visual references across scenes using reference-to-video inputs.
- Edit existing footage while preserving subject identity, motion, or continuity.
- Extend an existing clip for longer narrative sequences or commercial cuts.
- Produce e-commerce product videos for listings, PDPs, product pages, and advertisements.
- Create vertical short-form content for social commerce, livestream promotion, seeding, and UGC campaigns.
- Develop TVC concepts, branded films, product demonstrations, short dramas, and animated comic scenes.
- Compare available model versions, routing strategies, and real-time pricing before submission.
- Submit long-running jobs without waiting locally, then retrieve them later with a task ID.
How Video Generation and Editing Works
- The CLI reads the prompt, selected mode, media inputs, API key, routing preference, and optional model parameters.
- The skill validates that the supplied inputs match the selected capability, such as images for image-to-video or video for editing and extension.
- Reference images, videos, or audio are uploaded automatically when required, and the resulting media IDs are attached to the generation request.
- The client discovers or selects the appropriate publicModelId. Supported mappings include
t2v,i2v,r2v,edit, andextend. - The client queries the current model catalog and pricing snapshot, then applies
COST_FIRST,SPEED_FIRST, orSUCCESS_FIRSTrouting. - The request is submitted through the AI Hive OpenAPI, returning a
taskIdfor asynchronous processing. - The skill polls the original task rather than creating a duplicate request, which helps avoid unnecessary charges when local polling times out.
- When processing completes, the generated video is downloaded to the configured output directory unless
--no-downloadwas selected. - The resulting media and task information can be reviewed for production, iteration, comparison, or downstream editing.
Video Generation and Editing Setup
Requirements
- Python 3.6 or newer
- The
requestspackage - An AI Hive API key in the
sk-api-*format - The skill directory containing
scripts/videogen.py
Install the dependency
pip3 install requests
Initialize credentials
The recommended initialization flow opens AI Hive, guides you through login and API key creation, and stores the key in ~/.ai-hive/config.json with restricted permissions.
python3 "$SKILL_PATH/scripts/videogen.py" init --skill-name video-generation-and-editing
Alternatively, configure the key using an environment variable, command-line option, or configuration file:
export AI_HIVE_API_KEY=sk-api-your-key
python3 "$SKILL_PATH/scripts/videogen.py" generate --mode t2v --prompt "A cinematic product reveal" --no-download
Generate a first video
python3 "$SKILL_PATH/scripts/videogen.py" generate \
--mode t2v \
--prompt "A cinematic product commercial with a slow camera push-in and premium studio lighting"
Use existing media
python3 "$SKILL_PATH/scripts/videogen.py" generate \
--mode edit \
--prompt "Preserve the original subject and motion, change the environment to a premium commercial studio" \
--video /path/to/source.mp4
Submit and check later
python3 "$SKILL_PATH/scripts/videogen.py" generate --mode t2v --prompt "A complex cinematic scene" --no-download
python3 "$SKILL_PATH/scripts/videogen.py" task --task-id <taskId>
The default output location is ~/Downloads/AiHive. Use --output-dir to change it, and use --api-key, --base-url, or --verbose for runtime configuration and diagnostics.
Video Generation and Editing Data Schema & Taxonomy
Project structure
video-generation-and-editing/
├── SKILL.md
├── CHANGELOG.md
├── scripts/
│ └── videogen.py
└── resources/
└── config.example.json
Generation modes
| Mode | Capability | publicModelId |
|---|---|---|
t2v |
Text-to-video | public_model_seedance_2_5_t2v |
i2v |
Image-to-video | public_model_seedance_2_5_i2v |
r2v |
Reference-to-video | public_model_seedance_2_5_r2v |
edit |
Video editing | public_model_seedance_2_5_video_edit |
extend |
Video extension | public_model_seedance_2_5_video_extend |
Input and metadata taxonomy
prompt: Required natural-language description of the desired video.mode: Generation capability, includingt2v,i2v,r2v,edit, andextend.first-frameandlast-frame: Boundary images for controlled animation.image: One or more reference images.video: One or more source or reference videos.audio: One or more reference audio files.model-id: Explicit model selection when multiple matching candidates exist.routing: Cost, speed, or success preference.param: Runtime model settings inkey=valueform, such as resolution or duration.taskId: Identifier used to track asynchronous generation.mediaId: Identifier returned after reference media upload.pricingSnapshot: Runtime model and price information associated with submission.
Files and outputs
| Item | Default or format |
|---|---|
| Credentials | ~/.ai-hive/config.json, permissions 0600 |
| Generated output | ~/Downloads/AiHive |
| Video formats | MP4, MOV, or formats supported by the active model |
| Task state | Stored and queried through the returned taskId |
| Media assets | Uploaded through the AI Hive object-storage workflow and referenced by media IDs |
Video Generation and Editing Advanced Features
- Supports five specialized workflows: text-to-video, image-to-video, reference-to-video, video editing, and video extension.
- Accepts first-frame, last-frame, image, video, and audio controls for more constrained generation.
- Uses explicit
publicModelIdmappings for stable, predictable capability routing. - Queries real-time model availability and pricing instead of relying on hard-coded promotional prices.
- Provides
COST_FIRST,SPEED_FIRST, andSUCCESS_FIRSTrouting strategies. - Passes model-specific options through repeatable
--param key=valuearguments. - Supports asynchronous submission with
--no-download, allowing external scheduling or later task retrieval. - Automates upload credentials, object-storage uploads, and upload completion confirmation.
- Includes task polling and result downloading through the
tasksubcommand. - Offers verbose logging, custom API base URLs, and multiple API-key configuration methods.
- Enables sequential batch production for SKU variants, campaign assets, and multi-channel creative testing.
- Provides prompt guidance for separating subject motion from camera motion and specifying lighting, sound, constraints, and output format.
- Fits Openclaw Skills workflows that need repeatable AI video generation, content automation, and command-line integration.
SKILL.md
Loading
Related Openclaw Skills
Seedance 2.0 Discount Video Generation Channel
Create production-ready Seedance 2.0 videos from prompts and reference media with automated uploads, pricing checks, task tracking, and downloads.
Seedance 2.5 First-and-Last-Frame Video
Generate controlled Seedance 2.5 videos by defining the opening and closing frames, then automate media upload, task tracking, and result downloads.
Completion Report Generator
An automated Openclaw Skills report generator that aggregates task outcomes into daily tenant reports, fairness metrics, and 30-day welcome-back summaries.
回译反思法 · 越南语翻译实战版
A rigorous 10-stage Chinese-Vietnamese translation workflow that combines terminology management, multi-model reflection, blind back-translation, and human review.
Bazi Engine
Bazi Engine generates traceable Four Pillars charts and grounded traditional Chinese astrology reports from verified birth data.
Product Image Suite Generation & Editing
Generate and revise a fact-checked, channel-ready image suite for one SKU while preserving product identity, claims, and visual consistency.
Featured*
