HUANGCHIHHUNGLeo/claude-real-video

Most AI tools don't really see a video. Paste a YouTube link into ChatGPT and it reads the transcript, not the picture. Claude won't take a video file at all. Even Gemini, which can read video natively, has to send it up to Google and samples frames at a fixed interval (1 fps by default), so fast cuts slip past.

claude-real-video does it differently, and the processing runs locally: point it at a URL or a file, and it pulls the frames that actually matter (every scene change, not a fixed quota), throws away the near-duplicates, transcribes the audio, and hands you a clean folder any LLM can read. All the processing happens on your own machine — what gets sent anywhere is only the frames/text you choose to paste into an LLM afterwards.

crv "https://www.youtube.com/watch?v=..."
# → crv-out/frames/*.jpg  +  crv-out/transcript.txt (+ transcript.json with timestamps)  +  crv-out/MANIFEST.txt

Then drop the frames + MANIFEST.txt into Claude / ChatGPT / Gemini and ask away.

No terminal needed — run crv-web and a local page opens (Traditional Chinese / Simplified Chinese / English): paste a YouTube or Reels link or a file path, click Analyze, open the result viewer. Video analysis and output generation run on your machine — the source video never gets uploaded. (If you then paste the extracted frames or transcript into a cloud LLM, that data goes to that provider.)

Want to eyeball what the model will see first? Add --viewer — it writes a local viewer.html (video + keyframe grid + transcript) you can double-click open. No network, no extra installs.

Slow-changing content (animation tutorials, gradual morphs, slow pans): add --adaptive — frames are picked against their rolling neighbourhood instead of a fixed threshold, so a 2-3s squash-and-stretch that never spikes any single frame still gets captured.

Text-heavy content (lecture slides, screen recordings, talking-head explainers): add --text-anchors — extra frames are forced at subtitle-cue timestamps, so each spoken segment gets a matching visual even when the scene barely changes. Needs a sidecar .srt/.vtt or an embedded subtitle track — captions burned into the pixels can't be detected. At most one forced frame per second; scene detection is untouched.

Not doing LLM work? It also works as a general-purpose video keyframe extractor — scene-change detection + dedup, no ML models to download.

Using Claude Code? Install it as a skill so Claude watches videos on its own (the skills/ folder lives in the repo, not in the pip package — clone it first):

pip install "claude-real-video[whisper]"
git clone https://github.com/HUANGCHIHHUNGLeo/claude-real-video.git
mkdir -p ~/.claude/skills && cp -r claude-real-video/skills/claude-real-video ~/.claude/skills/

Then just paste a video link into Claude Code and ask about it.

New in 0.3.0 — tell it why you're watching, and keep what it finds:

crv "https://youtu.be/..." --why "find the pricing strategy" --kb ~/notes

--why makes the analysis focus on what you care about instead of a generic summary; --kb saves the result as a dated note in your own notes folder, so it doesn't die in crv-out.

Why not just sample frames?

Most "let an LLM watch a video" scripts (and Gemini's own pipeline) grab frames at a fixed interval — e.g. one per second. That over-samples a static screencast and under-samples a fast-cut reel. claude-real-video is smarter:

You feed the model fewer, more meaningful frames — cheaper context, better understanding.

Install

pip install "claude-real-video[whisper]"   # recommended: frames + dedup + audio transcription
pip install claude-real-video              # core only (frames + dedup)

pip extras never install themselves — without [whisper] there is no speech-to-text (videos that ship their own subtitles still get a transcript).

System requirement: ffmpeg

ffmpeg / ffprobe are used for frame extraction and audio, and aren't pip-installable. Install them once:

Verify it's on your PATH:

ffmpeg -version

Transcription uses the whisper CLI (installed by the [whisper] extra, or pip install openai-whisper). Whisper also relies on ffmpeg.

Works on macOS, Windows, and Linux — Python 3.10+.

Usage

# A YouTube / Instagram / TikTok / ... link
crv "https://www.instagram.com/reel/XXXX/"

# A local file, English transcript, output to ./out
crv lecture.mp4 -o out --lang en

# Frames only, no transcription
crv clip.mp4 --no-transcribe

# A login-gated video (your own / authorised use): pass a Netscape cookie file
crv "https://..." --cookies cookies.txt

python -m claude_real_video ... works as an alias for crv too.

Options

What --grid output looks like

One contact sheet = nine consecutive keyframes, in order, filenames on each cell — the model reads a sequence, not scattered stills:

Use it from Python

from claude_real_video import process

r = process("https://youtu.be/...", "out", lang="en")
print(r.frame_count, r.transcript_path)

How it works

1. Fetch — yt-dlp for URLs (optional cookies), or copy a local file.
2. Extract — one chronological ffmpeg select pass grabs every scene change plus a density floor (at least one frame every --fps-floor seconds), so fast cuts and slow screencasts are both covered.
3. Dedup — two detectors against a sliding window of the last --dedup-window kept frames, so an A-B-A cutaway doesn't re-send a shot the model has already seen. A global channel measures real pixel difference (downscaled RGB, not a perceptual hash — hashes go blind on flat colours and equal-luma hue changes); --dedup-threshold is the % of it that must change. A settled-local channel (v0.7.4) catches what the global one can't see: thin pen strokes, caption/text-card swaps and small UI updates that average out to ~0% globally. It looks, on a finer signature, for a region that differs strongly from every recent kept frame (with 1px shift tolerance, so film grain and frame jitter don't trigger) and is no longer changing — a settled new state, not motion mid-flight — with a cooldown so continuous motion that pauses every second (a waving flag, drifting smoke) can't keep re-firing. The final frame is evaluated even if still in motion (so a video's closing state is never lost), but it must clear both contrast gates like any other frame. --report writes report.html showing every keep/drop decision with its diff % (settled-local keeps are labelled), for tuning.
4. Text — if the video already has subtitles (a sidecar .srt/.vtt next to a local file, or an embedded subtitle track), those are used as the transcript — faster and more accurate than re-transcribing. Only when there are no subtitles does it fall back to Whisper on the audio (skipped cleanly if there's no audio).
5. Audio (optional, --keep-audio) — save the full original soundtrack (audio.m4a: music + speech + effects, copied losslessly when possible). The transcript only has the words; the audio file lets a model that can listen (Gemini, GPT-4o, …) actually hear the music and tone.
6. Manifest — MANIFEST.txt summarises everything for the model.

So the model can see (key frames), read (transcript) and — with --keep-audio — hear (full soundtrack) the video. The transcript is plain text any model can read; the tool doesn't burn subtitles into the video — burning is a presentation choice, not something needed to make a video AI-readable.

Notes

• Only download content you have the right to. The --cookies option is for your own, authorised access — don't ship credentials in a repo.
• Use one output folder per video. Re-running into a folder that already holds an analysis is refused (so two videos never mix); pass --overwrite to replace it.

crv Pro — understand how a video was shot

The free version tells your AI what's on screen. crv Pro tells it how it was shot — and why it works. Camera moves, editing rhythm, action bursts, plus a one-flag --breakdown report: hook analysis, pacing curve, camera language, Reels-algorithm lens, and a rubric your own LLM completes into a full video teardown.

This free tool tells an LLM what is on screen. A stack of keyframes can't tell it how the video moves — the camera work and the pacing.

crv Pro adds everything the free version can't hear or feel:

• Camera-move classification — every shot labelled static / pan / tilt / zoom / handheld (tested against a manually labeled internal sample set — public benchmark methodology is in progress)
• Editing rhythm — shot list, cuts per minute, and how pacing shifts across the video
• Perception timeline — the subtle things frames can't show: gestures and expressions (a smile, a hand raised, pointing), voice pitch rises and pauses, speaker emotion, and non-speech sound events — all timestamped
• A breakdown report — hook analysis, pacing curve, camera language, and a rubric your own LLM completes into a full teardown
• Three modes — --mode watch (understand the content), --mode creator (reverse-engineer the making), --mode full

Recent Pro updates (July 2026): a music-state timeline (hear the score building, peaking, falling away — with BPM), voice emotion read from the isolated voice instead of the full mix, an interactive --viewer dashboard with a clickable synced timeline, and richer gesture narration ("hand raised — right hand, while walking toward frame right").

All as plain text in the same manifest, all computed on your machine. One-time founder price $19:

• Buy on Capafy (instant download, license key included): https://capafy.ai/agent/llm-real-video-pro-let-any-llm-watch-videos/5451082151
• Product page & demo: https://leoaido.com/crv-pro/

Following the build? I'm documenting the road from open-source tool to first paying customer, in public — @LeoAidoAI on X.

License

MIT