OrkasVideoStudio skill
Skills, a CLI and an MCP server that let Claude Code or Codex make video from a brief: editing, graphics, voiceover, subtitles
- 331 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Runs ffmpeg and ovs commands on your machine and writes project files
- Generation uses your provider keys and spends money on every billable clip
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/orkasvideostudioDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Runs ffmpeg and ovs commands on your machine and writes project files
- Generation uses your provider keys and spends money on every billable clip
This is third-party code. Review the repository files before installing.
What it does
OrkasVideoStudio turns a coding agent into a video studio. You describe a video in plain words. The agent reads the material, writes a plan and produces the file. The plan lives in plan.json, so you can read and edit it, and changing one line re-renders only that piece. There are four lines of work. Compose builds motion graphics in HTML and renders them to mp4 with no paid keys. Edit cuts and joins your footage and trims silence and filler words. It also mixes audio, burns in subtitles and dubs voiceover. Generate pulls AI footage and images from providers on your keys. Auto weaves all three into one plan and checks before delivery that the result is not a slideshow. HyperFrames, ffmpeg and whisper.cpp do the work underneath. The agent first reads the video-router skill, which locks the line, then the stage skills such as stage-plan, stage-compose and video-craft. A generation budget is built into the rules: the skills try to avoid paid clips.
Who it is for. For developers and content makers who want to build explainers, promo videos and highlight clips from screen recordings right inside Claude Code, Codex or Cursor.
Good fit when
- You need a 60-second vertical clip with motion graphics, voiceover and subtitles and do not want to open a video editor
- You have an hour-long screen recording and need three short highlight clips with burned-in subtitles
- A webcam take is full of pauses and filler words: trim-silence and remove-fillers tighten it and leave an auditable trail of decisions
- A finished demo needs an English voiceover, subtitles and normalized loudness
- The video mixes sources: your footage, a generated opener and stats graphics in one plan
Not a fit when
- You need frame-by-frame manual editing with color grading and complex effects: the project works through an agent and commands and has no timeline
- The machine has no Node 22 and no ffmpeg with ffprobe, and you cannot install them
- You want a full video generated with one button by a hosted service: AI footage here goes only through your own provider keys
- You want a ready cloud service with an interface: work here happens in the agent's terminal
Example request
Make a 60-second vertical video explaining what a vector database is, with voiceover and subtitles. Show me the plan first.Limitations
- Needs Node 22 or newer, plus ffmpeg and ffprobe on PATH.
- Compose, edit and transcription work without keys. Generating images, video and voiceover needs your own keys: images through an OpenAI-compatible API, Gemini or Doubao Seedream, video through Doubao Seedance, Atlas Cloud or MuAPI, voiceover through an OpenAI-compatible TTS. The OpenAI and Gemini APIs are unavailable from Russia.
- Every generated scene is billed on your key. The skills cap a run at six generated shots and three recurring characters, but the project gives no price guarantee.
- The npm packages are not published yet, so the toolkit installs from source.
How to disable. Remove the MCP server with claude mcp remove ovs and delete the skill folders from ~/.claude/skills or ~/.agents/skills. The key config sits in ~/.config/orkas-video-studio/config.json; delete it if you no longer need the keys.
FAQ
How do I install OrkasVideoStudio in Claude Code?
Build the CLI from source: git clone, then `pnpm install` and `pnpm build`. The command `ovs skills --install --target claude` puts the skills into ~/.claude/skills. The MCP server from your build is added through claude mcp add with node and the packages/mcp/dist/index.js file, and once the packages are published `claude mcp add ovs -- npx -y @orkas/video-studio-mcp` will work. For Codex, use codex instead of claude.
Do I need paid keys to make a video?
Graphics, editing, transcription and subtitles need no keys. They are required only for generating images, video and voiceover through providers.
What is plan.json?
It is the plan of the video that the agent writes and you can edit. Each segment, caption and voiceover line re-renders on its own. The command ovs plan validate checks the plan, and the number of billable generations in it must equal the number of generate segments.
How is OrkasVideoStudio different from Remotion?
Remotion has you write video as React code. Here the agent decides which line fits, and HyperFrames, ffmpeg and whisper.cpp act as executors under a single plan.
How do I make sure the video is not just a slideshow?
The delivery guard handles it: `ovs plan promise-check plan.json` compares the finished cut with what the plan promised, for example real motion on screen. The agent runs it before handing over the file.
Which commands does ovs have?
Commands are grouped by line of work. Editing has subcommands for trimming, joining, burning in subtitles and mixing, and separate commands handle transcription, voiceover, the plan and skill installation. `ovs --help` prints the full list.
Discussion
No questions or reviews yet.
Related
SkillsA video editing skill for Claude Code and Codex: the agent cuts pauses and filler words, grades color, burns subtitles and adds animated overlays
A local video studio for your coding agent: describe a video in one sentence, the agent directs, your machine renders with no cloud
Agentic video studio for Claude Code, Cursor and Codex: script, voiceover, images, stock footage, music, subtitles and render
ChatGPT short-video editing skills
ChatGPT 剪短影音 Skills
Two agent skills: set up the editing environment and build a vertical short video in eight verified steps
A set of 11 skills for Claude Code and Codex: from an idea or a novel to episode scripts, storyboards, image and video prompts, and a finished cut
A skill for Claude Code and Codex that builds cinematic product videos with Remotion from shot recipe cards, a ready template and audio assets
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Repository stars | Main difference |
|---|---|---|---|
| OrkasVideoStudio (this entry) | no | 331 | |
| video-use | no | 30,108 | cuts pauses and filler words in finished footage, without generation or motion graphics |
| Claude Code Video Toolkit | no | 2,378 | a kit for script, voiceover, music and MP4 rendering in Claude Code |
| Remotion Agent Skills | no | 5,855 | official Remotion skills for video in React |
README in short
The project grew out of the built-in video agent of the Orkas app and is now a standalone MIT-licensed toolkit. For each generation provider the author lists the environment variables for its keys.
README badge
Original SKILL.md
--- name: video-router description: Read this FIRST on any video-production request to pick and lock the line — generate (AI footage), compose (designed HTML), edit (cut real footage), or AUTO end-to-end. Trigger when classifying a brief and committing to a primary path; do NOT trigger for authoring compositions, generating assets, or rendering output (those have their own stage skills). --- # video-router Knowledge for picking a video production line and locking it before work begins. This skill is read for guidance; it describes **what to decide**, not any tool mechanics.