videodb

Ingest video files or URLs and index spoken words for natural-language search.

Updated May 4, 2026
One-click install
npx skills add https://github.com/gganbukim1/myskills --skill videodb-gganbukim1
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: videodb
Source: https://github.com/gganbukim1/myskills/tree/main/videodb
Command: npx skills add https://github.com/gganbukim1/myskills --skill videodb-gganbukim1

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires python-dotenv, videodb, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Turning long or live video into searchable, actionable evidence without manually scrubbing timelines, while enabling real-time alerts from screen and audio context.

Core Features & Use Cases

  • Video ingest & streaming: Ingest files or URLs (including RTSP/live via RTStream/capture) and return playable stream links.
  • Perception & indexing: Build spoken-word, visual, and temporal indexes to locate moments with timestamps and evidence links.
  • Timeline edits & generation: Transcode/normalize media, generate subtitles and overlays, and compose clips into shareable streams.
  • Real-time monitoring & alerts: Capture desktop or ingest live streams, stream AI results via WebSocket, and emit event/alert payloads for detections.

Quick Start

Use the videodb skill to ingest a video URL, index spoken words, then search for a topic and compile the matching moments into a single playable stream.

Frequently Asked Questions about videodb

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I index spoken words in a video to search for specific moments?

To index spoken words in a video, ingest your media via files or URLs to build spoken-word and optional scene indexes. This enables natural-language search for relevant moments and returns timestamped evidence links for precise playback.

Can I capture live streams and receive real-time alerts based on video context?

Yes, you can ingest live streams or capture desktops to enable real-time monitoring. The system streams AI results via WebSocket and emits event or alert payloads for detected visual and spoken-word context.

Does the VideoDB API support timeline editing for clips and subtitles?

Yes, the VideoDB API supports timeline operations to transcode and normalize media. You can generate subtitles, apply overlays, and compose specific clips into shareable, playable stream URLs.

What is the best way to handle video ingestion for offline files and URL streams?

The best way to handle video ingestion is by processing offline video files or URL streams, including RTSP or live captures. This builds temporal indexes that locate moments with timestamps and returns playable stream links.

How do I compile matching video search results into a single playable stream?

After indexing your video and running a natural-language search for relevant moments, use timeline operations to compile the matching clips. This generates a single shareable stream URL for continuous playback.

What happens when a natural-language video search returns no results found?

When a natural-language video search returns no results found, the system safely handles the error. This prevents workflow interruptions during spoken-word or scene indexing and ensures your application returns a safe response.