Score breakdown
Popularity is tracked separately. Support, ads, sponsorships, and tips never affect these signals.
Why it matters
Useful for AI agent developers, video / media researchers, RAG teams, knowledge-base builders, content / e-learning creators, and accessibility engineers who need a local-first, scene-aware, deduplicated video frame + transcript pipeline that runs offline and exposes a Claude Code skill (`skills/claude-real-video/`) for drop-in install. It collapses a 10-minute static slide deck into 1 frame (vs ~
Who should use it
Who should skip it
Pass on HUANGCHIHHUNGLeo/claude-real-video if its scope or audience does not match what your team is building right now.
About this signal
HUANGCHIHHUNGLeo/claude-real-video is tracked by RepoRadar as a video frame extractor in the AI Agents section. It was first seen on 2026-07-03 and last updated on 2026-07-03. The current verdict is 'try now' with a Gold tier and easy setup difficulty. Across RepoRadar's eight signals, HUANGCHIHHUNGLeo/claude-real-video is strongest on workflow potential (9.8) and practical usefulness (9.0) and weakest on maturity (6.5) — a profile worth weighing against your own priorities. This page summarizes the evidence RepoRadar has captured from captured source metadata. The score, tier, risk label, and verdict on this page are never influenced by sponsorship, ads, or tips — they reflect only the usefulness, popularity, novelty, momentum, maturity, and evidence signals described in the RepoRadar methodology.
How this item is evaluated
RepoRadar assigned HUANGCHIHHUNGLeo/claude-real-video a composite score of 8.3 out of 10, placing it in the Gold tier. This score combines weighted sub-signals: usefulness (35%), novelty (18%), momentum (14%), maturity (10%), open-source/build quality (7%), evidence quality (6%), workflow potential (6%), and setup ease (4%). Popularity is tracked separately at 1.0 and never affects the composite score or tier. The risk label of 'conditional' reflects inherent user-impacting hazards, not generic novelty. Items with no risk flag may still require normal code review before production use.
Putting this into practice? Read How to vet an AI agent or MCP server before you wire it in for the checklist behind this score.
Risk explanation
Requires ffmpeg, opencv-python-headless, scenedetect, yt-dlp on PATH and a Python 3.10+ environment; on Windows, the bundled `install.ps1` handles PATH setup, on macOS the bundled `install.sh` handles Homebrew ffmpeg install, and on Linux the dependency install is manual (`apt install ffmpeg` then `pip install claude-real-video[whisper]`). Review the README's `System requirement: ffmpeg` section before promoting to multi-user deployment; The `--why` / `--kb` flags save dated notes to the user's notes folder — review the note path and the `MANIFEST.txt` before sharing the saved note across a team, because the saved note inherits any PII / NDA-sensitive content from the source video. The processing itself runs locally and only the user-chosen frames / text leaves the box, which is the right default for sensitive material; Whisper transcription quality depends on the model size and the audio quality — for noisy audio or accented speech, the smaller models under-transcribe, and the larger models are slower. The README documents the `faster-whisper` backend and the model size trade-off; pick the model size per use case (tiny / base / small / medium / large-v3).