AgentSkillsCN

video-gen

利用腾讯云VOD AIGC模型(Kling、Vidu、Hailuo等)生成视频。现已全面支持API调用。

SKILL.md
--- frontmatter
name: video-gen
description: Generate videos using Tencent Cloud VOD AIGC models (Kling, Vidu, Hailuo, etc.). Updated with full API support.

Video Generation Skill (VOD AIGC)

This skill generates high-quality AI videos using Tencent Cloud VOD AIGC API. It supports Text-to-Video (T2V) and Image-to-Video (I2V) across multiple models.

Supported Models

  • Kling (1.6, 2.0, 2.1, 2.5, 2.6, o1)
  • Hailuo (02, 2.3)
  • Vidu (q2, q2-turbo, q2-pro)
  • Jimeng (3.0pro)
  • Seedance (1.0-pro, 1.5-pro, etc.)
  • GV (3.1)
  • OS (2.0)

Usage

bash
node skills/video-gen/index.js "<Prompt>" [options]

Options

FlagDescriptionDefault
--model <name>Model name (Kling, Hailuo, Vidu, etc.)Kling
--model-version <ver>Model version (e.g. 2.1, 2.5)2.5
--image <url>Reference image URL for I2VNone
--last-frame <url>Last frame image URL (Kling 2.1, Vidu, GV)None
--resolution <res>720P, 1080P, 2K, 4K1080P
--ratio <ratio>16:9, 9:16, 1:116:9
--enhanceEnable prompt enhancementDisabled
--chat-id <id>Feishu chat ID for progress cardsNone

Examples

1. Basic Text-to-Video (Kling 2.5)

bash
node skills/video-gen/index.js "A cyberpunk city in rain" --model Kling --model-version 2.5 --chat-id "oc_..."

2. Image-to-Video (Hailuo 2.3)

bash
node skills/video-gen/index.js "Make the cat jump" --model Hailuo --model-version 2.3 --image "https://example.com/cat.jpg"

3. Start & End Frame (Kling 2.1)

bash
node skills/video-gen/index.js "Morph from day to night" --model Kling --model-version 2.1 --image "https://.../day.jpg" --last-frame "https://.../night.jpg"

Credentials

Requires .env variables:

  • VITE_VOD_SECRET_ID
  • VITE_VOD_SECRET_KEY
  • VITE_VOD_SUB_APP_ID