{
  "markdown": "<p align=\"center\">\n  <img src=\"static/video-use-banner.png\" alt=\"video-use\" width=\"100%\">\n</p>\n\n# video-use\n\nIntroducing **video-use** — edit videos with Claude Code. 100% open source.\n\nDrop raw footage in a folder, chat with Claude Code, get `final.mp4` back. Works for any content — talking heads, montages, tutorials, travel, interviews — without presets or menus.\n\nTry video-use in [Browser Use Cloud](https://cloud.browser-use.com/v4?utm_campaign=video-use-use-in-cloud&utm_source=github).\n\n## What it does\n\n- **Cuts out filler words** (`umm`, `uh`, false starts) and dead space between takes\n- **Auto color grades** every segment (warm cinematic, neutral punch, or any custom ffmpeg chain)\n- **30ms audio fades** at every cut so you never hear a pop\n- **Burns subtitles** in your style — 2-word UPPERCASE chunks by default, fully customizable\n- **Generates animation overlays** via [HyperFrames](https://github.com/heygen-com/hyperframes), [Remotion](https://www.remotion.dev/), [Manim](https://www.manim.community/), or PIL — spawned in parallel sub-agents, one per animation\n- **Self-evaluates the rendered output** at every cut boundary before showing you anything\n- **Persists session memory** in `project.md` so next week's session picks up where you left off\n\n## Setup prompt\n\nPaste into Claude Code, Codex, Hermes, Openclaw, or any agent with shell access:\n\n```text\nSet up https://github.com/browser-use/video-use for me.\n\nRead install.md first to install this repo, wire up ffmpeg, register the skill with whichever agent you're running under, and set up the ElevenLabs API key — ask me to paste it when you need it. Then read SKILL.md for daily usage, and always read helpers/ because that's where the editing scripts live. After install, don't transcribe anything on your own — just tell me it's ready and wait for me to drop footage into a folder.\n```\n\nThe agent handles the clone, dependencies, skill registration, and prompts you once for your ElevenLabs API key (grab one at [elevenlabs.io/app/settings/api-keys](https://elevenlabs.io/app/settings/api-keys)).\n\nThen point your agent at a folder of raw takes:\n\n```bash\ncd /path/to/your/videos\nclaude    # or codex, hermes, etc.\n```\n\nFor always-on editing from your own VPS or Telegram, run the agent through [Browser Use Box](https://browser-use.com/bux). [Watch the 15-second demo](https://www.tiktok.com/@browser_use/video/7639824093721758989).\n\nAnd in the session:\n\n> edit these into a launch video\n\nIt inventories the sources, proposes a strategy, waits for your OK, then produces `edit/final.mp4` next to your sources. All outputs live in `<videos_dir>/edit/` — the skill directory stays clean.\n\n## Manual install\n\nIf you'd rather do it by hand:\n\n```bash\n# 1. Clone and symlink into your agent's skills directory\ngit clone https://github.com/browser-use/video-use ~/Developer/video-use\nln -sfn ~/Developer/video-use ~/.claude/skills/video-use        # Claude Code\n# ln -sfn ~/Developer/video-use ~/.codex/skills/video-use       # Codex\n\n# 2. Install deps\ncd ~/Developer/video-use\nuv sync                         # or: pip install -e .\nbrew install ffmpeg             # required\nbrew install yt-dlp             # optional, for downloading online sources\n\n# 3. Add your ElevenLabs API key\ncp .env.example .env\n$EDITOR .env                    # ELEVENLABS_API_KEY=...\n```\n\n## How it works\n\nThe LLM never watches the video. It **reads** it — through two layers that together give it everything it needs to cut with word-boundary precision.\n\n<p align=\"center\">\n  <img src=\"static/timeline-view.svg\" alt=\"timeline_view composite — filmstrip + speaker track + waveform + word labels + silence-gap cut candidates\" width=\"100%\">\n</p>\n\n**Layer 1 — Audio transcript (always loaded).** One ElevenLabs Scribe call per source gives word-level timestamps, speaker diarization, and audio events (`(laughter)`, `(applause)`, `(sigh)`). All takes pack into a single ~12KB `takes_packed.md` — the LLM's primary reading view.\n\n```\n## C0103  (duration: 43.0s, 8 phrases)\n  [002.52-005.36] S0 Ninety percent of what a web agent does is completely wasted.\n  [006.08-006.74] S0 We fixed this.\n```\n\n**Layer 2 — Visual composite (on demand).** `timeline_view` produces a filmstrip + waveform + word labels PNG for any time range. Called only at decision points — ambiguous pauses, retake comparisons, cut-point sanity checks.\n\n> Naive approach: 30,000 frames × 1,500 tokens = **45M tokens of noise**.\n> Video Use: **12KB text + a handful of PNGs**.\n\nSame idea as browser-use giving an LLM a structured DOM instead of a screenshot — but for video.\n\n## Pipeline\n\n```\nTranscribe ──> Pack ──> LLM Reasons ──> EDL ──> Render ──> Self-Eval\n                                                              │\n                                                              └─ issue? fix + re-render (max 3)\n```\n\nThe self-eval loop runs `timeline_view` on the _rendered output_ at every cut boundary — catches visual jumps, audio pops, hidden subtitles. You see the preview only after it passes.\n\n## Design principles\n\n1. **Text + on-demand visuals.** No frame-dumping. The transcript is the surface.\n2. **Audio is primary, visuals follow.** Cuts come from speech boundaries and silence gaps.\n3. **Ask → confirm → execute → self-eval → persist.** Never touch the cut without strategy approval.\n4. **Zero assumptions about content type.** Look, ask, then edit.\n5. **12 hard rules, artistic freedom elsewhere.** Production-correctness is non-negotiable. Taste isn't.\n\nSee [`SKILL.md`](./SKILL.md) for the full production rules and editing craft.\n",
  "bytes": 5578,
  "sha": "612bca23557cfc4501e80d3ce4babbe086dba77a70fa3a74ae5c0170393ece8f",
  "repo_slug": "browser-use/video-use",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/skl_browser_use_video_use_video_use_ca8d40fe/readme"
}