Otopilot workflow
A local short-video studio: topic, script, CC0 images, voice-over and MP4 with subtitles, driven by Claude Code or Codex
- 245 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- The panel and scripts run ffmpeg and network requests to image and voice sources on your machine
- Script writing and voice-over use API keys and paid services, spending money on your key
- The unofficial Edge voice may stop working
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add workflows/otopilotDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- The panel and scripts run ffmpeg and network requests to image and voice sources on your machine
- Script writing and voice-over use API keys and paid services, spending money on your key
- The unofficial Edge voice may stop working
This is third-party code. Review the repository files before installing.
What it does
Otopilot is an open-source tool that runs on your own computer: a topic becomes a script, free images and CC0 museum art, voice-over, an edit, a finished MP4 with subtitles, a thumbnail and a YouTube pack. It also builds short series with characters and dialogue. You hand the whole repository to Claude Code or Codex, and the agent takes install and commands from `AGENTS.md`. The script is written by Claude or vidIQ with your key, or you paste your own, and everything else works with no keys. Image sources: Wikimedia, Openverse, NASA and museums (Met, Art Institute of Chicago, Cleveland, Rijksmuseum), optionally Pexels, Pixabay, Unsplash, Smithsonian and Europeana. Voice-over uses the free Edge voice (unofficial) with a fallback to the system voice, or Azure Speech, Piper and VoxCPM2. ffmpeg does the edit and loudness is normalized to −16 LUFS. The panel opens on `localhost:4173`. Formats: YouTube long form 16:9, Shorts, Reels and TikTok. Before generating, the panel shows a cost estimate and keeps a spend log, and there is no automatic YouTube upload: publishing stays yours.
Who it is for. Short video creators and knowledge channels who want to assemble clips locally with an agent, with no editing software or subscription service.
Good fit when
- A topic and a title should turn into a 40-second Shorts video with CC0 images and voice-over
- You want a series with characters and dialogue the agent assembles from a brief
- Every image needs CC BY attribution in the publishing pack
- Before starting you want a cost estimate and a daily limit
- Vertical clips must be cut from a long video
Not a fit when
- You want a cloud service that takes video orders: the project runs on your computer only and makes no videos to order
- You need automatic YouTube upload: there is none and publishing stays manual
- You need stable work on macOS or Linux: Windows is the main platform and the others are lightly tested
Example request
Make a 40-second Shorts video about the Rijksmuseum with CC0 images and voice-over, showing the script for approval firstLimitations
It needs Node.js 22 or newer and ffmpeg. Windows is the main platform, while macOS and Linux are called experimental. The model-written script needs a Claude or vidIQ key, and the free Edge voice is unofficial and may stop working, in which case the system voice is used. Images under NC and ND licenses are filtered out, and CC BY attribution lands in the publishing pack. The README is mostly in Turkish with an English intro. The last commit is from November 2026.
How to disable. Delete the cloned otopilot folder along with uretim/ holding finished videos and the .env file with keys. Stop the node panel.js process.
FAQ
Are keys needed?
Not for the free path: Wikimedia, Openverse, NASA and museum images, the Edge voice and ffmpeg work with no keys. Script writing through Claude or vidIQ is the paid step, or you paste a script by hand.
How do I run it through Claude Code or Codex?
Open the repository as a folder and start the agent. It reads AGENTS.md itself (through CLAUDE.md for Claude Code), knows install and commands, and takes a request like make a 40-second Shorts.
Does it upload to YouTube?
No, there is deliberately no automatic upload. The tool prepares the title, description, chapters, tags, attribution and thumbnail, and publishing stays yours.
Which formats does it support?
YouTube long form 16:9, Shorts, Instagram Reels and TikTok in 9:16.
How do I track costs?
An estimate is shown before starting, spend is logged in uretim/<job>/kullanim.json, and limits can be set per day, month and job.
Discussion
No questions or reviews yet.
Related
WorkflowsA skill and CLI: the agent writes a script and scenes, while the voice, word timing and MP4 render run on your computer with no API keys
A local video studio for your coding agent: describe a video in one sentence, the agent directs, your machine renders with no cloud
Free Claude Code skills for short-form video: scripts, local editing, translation with a cloned voice and publishing.
Agentic video studio for Claude Code, Cursor and Codex: script, voiceover, images, stock footage, music, subtitles and render
Skills, a CLI and an MCP server that let Claude Code or Codex make video from a brief: editing, graphics, voiceover, subtitles
A skill and CLI that let Claude Code or Codex make videos as web pages, with the Studio editor to change the result
README in short
The description, in Turkish and English: a topic goes in, a finished video comes out. It then covers the pipeline from script to edit, formats, an image license filter, a publishing pack, cost tracking, crash resilience and install on Windows, macOS and Linux.