video-upscaler

The Video Upscaler skill provides professional-grade video quality enhancement by leveraging a powerful, multi-model backend. It intelligently selects the best AI model (Topaz, SeedVR2, etc.) based on the user-defined profile to achieve optimal results, transforming low-resolution or noisy footage into crisp, cinematic-quality video.

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "video-upscaler" with this command: npx skills add wells1137/media-skills/wells1137-media-skills-video-upscaler

Summary

The Video Upscaler skill provides professional-grade video quality enhancement by leveraging a powerful, multi-model backend. It intelligently selects the best AI model (Topaz, SeedVR2, etc.) based on the user-defined profile to achieve optimal results, transforming low-resolution or noisy footage into crisp, cinematic-quality video.

This skill abstracts away the complexity of choosing and configuring different AI upscaling models. Instead of dealing with dozens of technical parameters, the user simply chooses a high-level goal, and the skill handles the rest.

Features

  • Multi-Model Backend: Dynamically routes requests to the best model for the job (Topaz, SeedVR2, etc.) via a unified API.

  • Profile-Based Enhancement: Offers a range of pre-configured profiles for common use cases, from standard 2x upscaling to 4K cinematic conversion and 60 FPS frame boosting.

  • Asynchronous by Design: Handles long-running video processing jobs without blocking the agent.

  • Simple Interface: Requires only a video URL and a profile name to start.

How It Works

The skill operates in a simple, two-step asynchronous workflow:

Submit Job: The agent calls the /upscale endpoint with a video URL and a profile name. The service validates the request, selects the appropriate AI model, and submits the job to the fal.ai backend. It immediately returns a task_id .

Poll for Status: The agent uses the task_id to periodically call the /status/{task_id} endpoint. The status will be queued , in_progress , or completed . Once completed, the response will contain the URL of the final, upscaled video.

Available Profiles

Profile Name Description

standard_x2

2x upscale using Topaz Proteus v4. Best all-around quality for live-action footage.

cinema_4k

Upscale to 4K (2160p) using SeedVR2. Best for cinematic content requiring temporal consistency.

frame_boost_60fps

2x upscale + frame interpolation to 60 FPS using Topaz Apollo v8. Best for sports and action.

ai_video_enhance

4x upscale using Topaz. Best for AI-generated videos that need resolution boosting.

web_optimized

Upscale to 1080p with web-optimized H264 output. Best for social media and web publishing.

End-to-End Example

User Request: "Enhance this video to 4K cinematic quality: [video_url]"

  1. Agent -> Skill (Submit Job)

The agent identifies the user's intent and calls the /upscale endpoint with the cinema_4k profile.

curl -X POST http://<your_backend_url>/upscale
-H "Content-Type: application/json"
-d "video_url": "[video_url]", "profile": "cinema_4k" }

Response:

{ "task_id": "a1b2c3d4-e5f6-7890-1234-567890abcdef", "model_used": "fal-ai/seedvr/upscale/video", "profile": "cinema_4k" }

  1. Agent -> Skill (Poll for Status)

The agent waits and then polls the status endpoint.

curl http://<your_backend_url>/status/a1b2c3d4-e5f6-7890-1234-567890abcdef

Response (In Progress):

{ "task_id": "a1b2c3d4-e5f6-7890-1234-567890abcdef", "status": "in_progress", "logs": ["Processing frame 100/1200..."] }

Response (Completed):

{ "task_id": "a1b2c3d4-e5f6-7890-1234-567890abcdef", "status": "completed", "result": { "video_url": "https://.../upscaled_video.mp4" } }

  1. Agent -> User

The agent delivers the final, upscaled video URL to the user.

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

General

audiomind

No summary provided by upstream source.

Repository SourceNeeds Review
General

kling-studio

No summary provided by upstream source.

Repository SourceNeeds Review
General

video-overlay

No summary provided by upstream source.

Repository SourceNeeds Review
General

seedance-prompter

No summary provided by upstream source.

Repository SourceNeeds Review