Agnes Video Generator

Generates AI videos from text or images using Agnes-Video-V2.0 with customizable frames, resolution, and asynchronous task handling.

A skill for generating AI videos using Agnes-Video-V2.0 model.

Capabilities

• Text-to-video generation
• Image-to-video animation
• Multi-image and keyframe animation (requires script extension)
• Asynchronous task handling with auto-polling
• Returns direct video URL

Usage

The skill is invoked automatically when the user requests video generation. It accepts:

• prompt (string, required): Video description
• image (string, optional): Source image URL for img2vid
• num_frames (number, optional): Frame count, must be 8n+1 (e.g., 81, 121, 241, 441)
• frame_rate (number, optional): FPS, default 24
• size (string, optional): Resolution like "1152x768"
• seed (number, optional): Random seed
• negative_prompt (string, optional): What to avoid
• extra_mode (string, optional): e.g., "keyframes" for multi-image

Implementation

• Endpoint: https://apihub.agnes-ai.com/v1/videos
• Query: GET https://apihub.agnes-ai.com/agnesapi?video_id=...
• Authentication: Bearer token from AGNES_API_KEY environment variable
• Model: agnes-video-v2.0
• Polls every 5 seconds (up to ~10 minutes) until completion

Notes

• Video generation is asynchronous; script handles polling automatically.
• remixed_from_video_id field in the final response holds the video URL.
• Pricing: $0.005 per second (check Agnes docs for current rates).

Example

bash

node generate.js --prompt "A cinematic shot of a cat walking on the beach at sunset" --num_frames 121 --frame_rate 24 --size 1152x768