skills.rest
vellum-ai
vellum-aiOFFICIAL
Explore all skills by this author →
Website𝕏@vellum_aiGitHub

media-processing

Ingest media assets, deduplicate by content hash, and run a 3-phase pipeline.

1.0k|158|Updated Feb 7, 2026
Search Tags:#transcription#pipeline#ingest#media-processing#gemini-map#claude-reduce
One-click install
npx skills add https://github.com/vellum-ai/vellum-assistant --skill media-processing-vellum-ai
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: media-processing
Source: https://github.com/vellum-ai/vellum-assistant/tree/main/assistant/src/config/bundled-skills/media-processing
Command: npx skills add https://github.com/vellum-ai/vellum-assistant --skill media-processing-vellum-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Ingests media files and orchestrates automated processing (ingest, preprocess, map, reduce) to extract structured outputs for downstream analysis and interaction.

Core Features & Use Cases

  • Ingest, deduplicate by content hash, and track assets through a deterministic three-phase pipeline.
  • Preprocess: extract keyframes and optional transcripts; Map: Gemini-driven vision analysis with structured JSON outputs; Reduce: Claude-driven QA over structured results.
  • Clip generation and on-demand access to frames, transcripts, and metadata for reviews and reporting.

Quick Start

Upload a video to begin automated ingestion and pipeline processing.

Frequently Asked Questions about media-processing

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

FAQPage Schema
How do I automate video transcription and frame extraction in a single pipeline?▼

Automated video transcription and frame extraction is handled by a deterministic three-phase pipeline that preprocesses media assets, extracts keyframes, and generates optional transcripts during the initial ingest phase.

Can I use Claude for Q&A and Gemini for vision analysis on the same media assets?▼

Claude and Gemini can be used together on the same media assets through a map-reduce pipeline where Gemini drives vision analysis to produce structured JSON, and Claude performs subsequent Q&A over those structured results.

What is the best way to prevent duplicate media processing and track costs per segment?▼

Prevent duplicate processing by deduplicating assets through content hashing and enforcing idempotent reprocessing, while tracking costs and maintaining resumable progress through modular services with per-segment caching.

How do I extract structured JSON outputs from audio and image files for downstream analysis?▼

Extract structured JSON outputs from audio and image files by ingesting them into the pipeline, where Gemini-driven vision analysis processes visual content and audio transcription converts speech into tracked text metadata.

Does the media processing pipeline support resumable progress if a job fails midway?▼

The media processing pipeline supports resumable progress through per-segment caching and idempotent reprocessing, allowing interrupted jobs to resume without duplicating work or losing previously generated structured outputs.

Related Skills

Explore More Software Engineering Skills→
WUIX69WUIX69

fal-ai-media

Generates AI images, videos, and audio from natural-language prompts via fal.ai MCP.

Community
Advanced
jayhjenkinsjayhjenkins
1

meeting-synthesis

Extract actionable product signals from meeting transcripts with structured outputs.

Community
Basic
levanminhduclevanminhduc

gemini-video-understanding

Analyze local and YouTube videos with the Gemini API for summarization and Q&A.

Community
Advanced
video-production-buddyvideo-production-buddy
602

video-understand

Analyze local videos with ffmpeg frame extraction and Whisper transcription.

Official
Intermediate