Skip to content
Sign in
  • Browse all toolsFilter by category and pricing
  • Browse categoriesWriting, code, image, audio and more
  • Compare toolsPut up to three side by side
  • Read the blogGuides and comparisons

claude-video-vision

Claude Code plugin for video frame extraction, timestamped audio transcription, and YouTube video analysis.

claude-video-vision is a Claude Code plugin that extracts video frames and timestamped audio transcriptions for Claude to examine. Designed for Claude Code users working with local videos and YouTube URLs, the MIT-licensed project supports free local Whisper processing, a Gemini API free tier, and usage-paid OpenAI audio processing.

Updated 9 Nov 20266 sections

At a glance

Pricing
FreeMIT-licensed open-source plugin. Local Whisper processing is free; Gemini API has a free tier, and OpenAI API audio processing is paid per usage.
Added
9 Nov 2026
claude-video-vision website screenshot

What is claude-video-vision?

claude-video-vision is an open-source Claude Code plugin that gives Claude access to video frames and timestamped audio transcriptions. Claude Code users can supply local video files or YouTube URLs and ask questions about their contents. The repository identifies the plugin as a perception layer, not an interpretation layer, because Claude receives the extracted material and handles the interpretation.

claude-video-vision combines ffmpeg frame extraction with a choice of audio-processing backends. The project is distributed under the MIT license and includes an MCP server, a video-perception skill, and slash commands. This structure connects video processing to an existing Claude Code session rather than providing a separate video-viewing application.

What claude-video-vision can do

claude-video-vision provides visual and audio inputs for questions about video content. Frame extraction remains available across all supported audio backends, so Claude receives direct visual access regardless of the transcription method. The plugin also exposes tools for checking metadata, examining video structure before extraction, and returning to specific moments for closer inspection.

  • claude-video-vision extracts video frames through ffmpeg.
  • claude-video-vision supplies audio transcriptions with timestamps.
  • claude-video-vision downloads YouTube videos through yt-dlp while preserving source metadata and captions.
  • claude-video-vision lets Claude adjust frame rate, time range, and resolution.
  • claude-video-vision retrieves video metadata without processing the video.
  • claude-video-vision supports detailed inspection of cached or newly extracted moments.

How claude-video-vision works

claude-video-vision installs through the Claude Code plugin marketplace, with its MCP server automatically installed from npm through npx on first use. A local repository checkout is also supported. The configuration wizard guides backend selection, local Whisper settings, frame options, and dependency verification. Required dependencies include Node.js 24+ and ffmpeg; YouTube URLs additionally require yt-dlp.

claude-video-vision accepts a video path or YouTube URL through /watch-video, optionally followed by a question. Conversational references to videos are also supported. For YouTube transcripts, the server prefers available manual English subtitles, then automatic captions, and falls back to the configured audio backend when captions are missing, empty, or insufficient. Results identify the transcription source.

Who claude-video-vision is for

claude-video-vision serves Claude Code users who want to ask questions about recorded visual content without leaving their session. The documented examples cover tutorials, lectures, YouTube Shorts, and video bug reports. Questions can concern the full recording or a narrow moment, allowing Claude to adapt extraction settings to the requested task.

  • claude-video-vision supports summarizing a lecture or YouTube video.
  • claude-video-vision supports identifying the language used in a tutorial.
  • claude-video-vision supports inspecting the opening moment of a video bug report.
  • claude-video-vision supports reading on-screen text at a specified timestamp.

claude-video-vision pricing and access

claude-video-vision is an MIT-licensed open-source project available through its GitHub repository and the Claude Code plugin installation workflow. Audio-processing costs depend on the selected backend rather than a listed plugin subscription. Local Whisper processing is described as free and fully offline, while the cloud options use API credentials.

claude-video-vision lists a Gemini API free tier of 1500 requests per day and describes OpenAI Whisper API processing as paid per usage. Local processing supports whisper.cpp or Python openai-whisper, with Whisper models downloaded automatically on first use. These backend terms concern audio processing; the repository does not provide pricing for Claude Code access.

claude-video-vision strengths and limitations

claude-video-vision combines direct frame access, timestamped transcripts, configurable extraction, and multiple audio backends in a Claude Code workflow. Local Whisper offers an offline audio-processing option, while YouTube transcript provenance distinguishes manual subtitles from automatic captions. Setup assistance and automatic model downloads reduce some configuration steps, but the plugin still depends on external software and the selected backend.

  • claude-video-vision requires Claude Code, Node.js 24+, and ffmpeg.
  • claude-video-vision requires yt-dlp for YouTube URL processing.
  • claude-video-vision requires API keys for Gemini API or OpenAI API audio processing.
  • claude-video-vision provides perception inputs rather than independent video interpretation.

claude-video-vision FAQ

What is claude-video-vision?

claude-video-vision is an open-source Claude Code plugin that extracts video frames and processes audio into timestamped transcriptions. Claude receives these inputs to answer questions about local video files or YouTube content. The project describes the plugin as a perception layer rather than an interpretation layer.

Is claude-video-vision free?

claude-video-vision is available as an MIT-licensed open-source project, with free local Whisper audio processing. The Gemini API backend offers a free tier, while the OpenAI API backend is paid per usage. Backend charges are separate from the plugin's open-source availability.

Does claude-video-vision support YouTube videos?

claude-video-vision supports YouTube URLs through yt-dlp, preserving source metadata and captions. The server prefers manual English subtitles, then automatic captions, and uses the configured audio backend when captions are unavailable or insufficient. Audio results label the transcription source.

What does claude-video-vision require?

claude-video-vision requires Claude Code, Node.js 24+, and ffmpeg, with yt-dlp additionally required for YouTube URLs. Gemini and OpenAI audio backends require API keys. Local audio processing requires whisper.cpp or Python openai-whisper, and Whisper models download automatically on first use.

Can claude-video-vision process audio offline?

claude-video-vision supports free, fully offline audio processing through its local Whisper backend using whisper.cpp or Python openai-whisper. Whisper models download automatically on first use. This offline capability applies to the local audio backend, while the plugin supplies the resulting transcripts and video frames to Claude Code.

How do you use claude-video-vision in Claude Code?

claude-video-vision accepts local video paths and YouTube URLs through the /watch-video command, with an optional question. Claude Code users can also mention a video conversationally. The setup wizard guides backend selection, frame configuration, and dependency verification after plugin installation.

From the community

No reviews yet

Be the first to review this tool. What you write helps the next person choose.

More tools like claude-video-vision

Claude Video

www.github.com

Give Claude Code and other agents video evidence from URLs or local files, with timestamped frames and transcripts.

Free

claude-watch

www.github.com

Analyze videos in Claude with scene-change frames, timestamped transcripts, structured reports, and Obsidian saving.

Free

HyperFrames

www.github.com

Open-source HTML-to-video framework with local rendering, browser previews, and AI coding agent skills.

Freemium

orc-claude-plugin

www.github.com

Connect Claude to ORC8R Cloud through MCP to find operations, inspect inputs, and report results.

Free