Video2context — AI skill for Claude Code
Turn screen recordings into compact, timestamped context for AI coding agents.
How to install Video2context
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open pranavchauhann/video2context and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Video2context does
Turn screen recordings into compact, timestamped context for AI coding agents. A Python CLI that extracts key screenshots, optional speech and OCR, and exports Markdown + JSON for Claude Code, Codex, and Cursor. Local processing by default.
Alternatives in AI
- Peekaboo — Peekaboo is a macOS CLI & optional MCP server that enables AI agents to capture screenshots of applications, o 3.2k ★
- Azure Vision Java — Image captioning, OCR, and object detection 1.8k ★
- Claude Skill Aso Appstore Screenshots — A Claude Code skill that generates high-converting App Store screenshots for your iOS app 1.2k ★
README
Video2Context
Give your coding agent the context inside your video.
Turn screen recordings into **selected screenshots, timestamped evidence, and readable context** for Claude Code, Codex, Cursor, and other coding agents.
[](pyproject.toml) [](#quick-start) [](#privacy-and-limits) [](LICENSE)
**Video → useful evidence → your agent → your task**
[Quick start](#quick-start) · [New user](#new-user-setup-and-first-video) · [Existing user](#existing-user-process-another-video) · [Cleanup](#finished-delete-the-generated-files) · [Agent prompts](#use-with-your-agent) · [Providers](#speech-ocr-and-vision) · [Troubleshooting](#troubleshooting) · [Developer docs](#development)
What does it do?
A screen recording contains useful information scattered across repeated frames, spoken feedback, and on-screen text. Video2Context selects visual evidence and packages it so your agent can inspect it without consuming every video frame.
| Your recording | What you can ask your agent to do |
|---|---|
| UI feedback | Identify requested layout, content, or component changes |
| Bug reproduction | Trace actions, inspect visible errors, and investigate the issue |
| App walkthrough | Understand screens and behavior before implementation |
| Screencast of any app or website | Review the actions taken and explain possible mistakes |
**Video2Context prepares the evidence. You give the agent the task.** It does not run your agent, edit your code, or decide what you intended to do.
Your video Video2Context Your agent
────────── ───────────── ──────────
recording.mov ─
Related Skills
Ghost Showcase
Invisible AI assistant for Windows — real-time STT + multi-provider LLM (Claude/GPT/Ollama/LM Studio) + RAG, s
AI Video Transcript Keyframes
Turn any video into a timestamped transcript + labelled keyframe contact sheets so Claude, ChatGPT or Cursor c
Kankan Shoucang
专治收藏夹吃灰 —— 让存过的小红书笔记这次真的找得回来。不是爬虫,不用你的账号登录态发请求,封控风险低。拖拽导入 + 图片 OCR + 自动分类,全程本地,不需要 AI API key。macOS 桌面 App。
Thruuu Claude Content Strategist
Turn keyword cluster exports from thruuu into prioritized content strategies using Claude Code. This is not a
Codenanny
Turn AI-generated project sprawl into a searchable, contextualized library. Ingests Claude Code transcripts; s
Yt Video Analyzer
A Claude Code skill that turns a YouTube video into a digest with timestamps and interface screenshots. Subtit
Related Agents
Broll Worker
Cheap, narrow B-roll worker. Executes specific footage tasks from a B-roll GOD plan (stock video, images, clip
Observer
Visual analysis specialist. Interprets images, screenshots, PDFs, and diagrams; extracts exact text via OCR (n
OCR Reader
Batch OCR re-reader for low-confidence book pages. Reads page screenshot images via multimodal Read tool and e