# HUANGCHIHHUNGLeo/claude-real-video

Let Claude (or any LLM) actually watch a video — scene-aware, deduplicated frames + transcript, from a URL or local file. Runs locally, MIT.

Repository: https://github.com/HUANGCHIHHUNGLeo/claude-real-video
Canonical: https://ross.abutalabs.com/products/claude-real-video
Language: Python
License: MIT
License Family: permissive
Topics: claude, claude-code, cli, ffmpeg, keyframe-extraction, llm, multimodal, python, scene-detection, transcription, video-analysis, whisper
Last push: 2026-08-22T11:12:23+00:00

## Health v2 (maintenance only)
Score: 80/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 98, longevity 4
- inputs: {"age_days": 64, "days_push": 11, "days_rel": 15, "gap_med": 0, "n_releases_24m": 24}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2072, forks 180 (observed 2026-08-28T04:06:10.398086+00:00)

## What it is
A Python CLI tool and agent skill that lets LLMs like Claude actually watch videos by extracting scene-aware, deduplicated keyframes plus a transcript from a URL or local file. It integrates with Claude Code, Cursor, and other agent hosts via npx skills or a plugin marketplace.

## Use cases
- let claude watch a youtube video
- summarize a video with an llm
- extract keyframes from a video for ai analysis
- get a transcript of a video file
- ask questions about a video's visual content
- deduplicate video frames to save tokens
- analyze video content with a multimodal agent

## When to choose
- you want an LLM to reason about video visuals, not just the transcript
- you use Claude Code, Cursor, or another agent host and want video understanding as a skill
- you need local, MIT-licensed video frame extraction and transcription
- you want token-efficient frame sampling with contact sheets

## When to avoid
- you need a hosted video-understanding API with no local setup
- you only need video transcripts without visual analysis
- you need advanced cinematic analysis like cut rhythm and emotion timelines, which is behind the paid Pro add-on

## Facets
- artifact type: cli-tool
- maturity: active
- function: video-processing, cli, speech-recognition, machine-learning, developer-tools
- domain: artificial-intelligence, large-language-models, developer-tools
- platform: python, cli, cross-platform
- tags: video-analysis, keyframe-extraction, scene-detection, transcription, whisper, ffmpeg, multimodal, claude-code, llm-tools, agent-skills, video, command-line

## Member repositories
- HUANGCHIHHUNGLeo/claude-real-video (main) score 80

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:10.398086+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:56:40.650902+00:00, confidence not recorded.
  - readme: https://github.com/HUANGCHIHHUNGLeo/claude-real-video (fetched 2026-08-28T04:06:10.398086+00:00, sha 550d9f8479d4)
  - registry_pypi: https://pypi.org/pypi/claude-real-video/json (fetched 2026-08-29T10:36:45.064401+00:00, sha 85c98994b2f0)
- Data as of 2026-08-30T08:39:29.467469+00:00.