What open-source YouTube automation tools can and cannot do
Each tool here saves time on one or more steps of making a video. None of them can decide what your audience wants, and none of them change YouTube’s rules. Most need some setup: Python or Docker, an API key for a language model such as OpenAI or Gemini, and often a text-to-speech service. Expect to spend an afternoon getting your first video out of any of them.
Topic-to-video generators for faceless channels
These tools take a topic or a script and assemble a finished short video from stock footage, an AI voice, subtitles and music. They are the starting point for faceless YouTube automation, where the creator does not appear on camera.
MoneyPrinterTurbo is the most complete of the group. You give it a topic or keyword, and it writes the script, finds stock clips, adds a voiceover, subtitles and background music, then renders a video in portrait or landscape. It has a web interface and an API, and it is MIT licensed. Start here if you want to see what faceless video AI can do with the least code.
MoneyPrinter is a smaller MIT-licensed Python script that builds Shorts from a topic using a language model script, stock clips, text-to-speech and burned-in captions. It is easier to read and modify than MoneyPrinterTurbo, which makes it a good base for your own pipeline.
short-video-factory is a desktop app under AGPL-3.0 that generates short videos with AI scripts, stock footage, voice and subtitles. Pick it if you prefer clicking through an app to working in a terminal.
ShortGPT is an MIT-licensed framework that covers scripts, voice, editing, translation and dubbing for Shorts and TikTok. It has gone longer without updates than the tools above, so test it before you build on it.
short-video-maker creates vertical videos from text through an MCP server (Model Context Protocol, a standard way for AI assistants to call tools) and a REST API. It suits people who want an AI assistant or a workflow tool to request videos automatically.
AI-Faceless-Video-Generator writes a script, generates a voice and animates a talking face, so you get a presenter without filming anyone. RedditVideoMakerBot turns Reddit threads into narrated videos over background footage under GPL-3.0. Videos built from other people’s posts raise questions about permission and originality, so think carefully before you publish them.
Agents that automate a YouTube channel end to end
If you want to automate YouTube channel tasks beyond a single video, these projects chain the steps together.
youtube-automation-agent is a self-hosted AI agent that researches topics, writes scripts, creates narration and visuals, prepares titles and descriptions, and publishes after you approve each video. It works with several model providers and is MIT licensed. The approval step is the main reason to choose it, because you stay in control of what goes live.
youtube-shorts-pipeline turns news items into Shorts with a script, AI visuals, voiceover, captions and upload. It is a compact MIT-licensed example of AI YouTube automation that you can read in an evening and adapt.
AI shorts generator tools for long videos
If you already publish long videos or podcasts, clipping them into Shorts is often more useful than generating new videos from scratch, and these YouTube automation tools work from footage you already own.
OpenShorts finds strong moments in a long video and exports 9:16 clips with face tracking, subtitles and dubbing. It runs with Docker, is MIT licensed and includes an MCP server for AI agents.
AI-Youtube-Shorts-Generator uses a language model to pick highlights, Whisper (an open speech recognition model) for transcripts and face detection for vertical cropping. It is a good free AI shorts generator if you are comfortable running Python.
AutoClip is a desktop tool that cuts podcasts, interviews and courses into short clips with subtitles and cover images. yt-short-clipper is a Windows app that does a similar job with word-by-word captions. Both are MIT licensed and suit people who want a desktop app instead of a command line.
NarratoAI uses a language model to write commentary for a video, voices it and cuts the footage into a narrated recap. It is MIT licensed, and its project description is written in Chinese. Recaps of films or shows that you do not own raise copyright problems, so use it on footage you have the rights to.
Subtitles, translation and dubbing
VideoLingo subtitles, translates, aligns and dubs a video so you can publish it in other languages. It is Apache-2.0 licensed and is the most useful tool here for reaching viewers who speak another language.
VideoCaptioner is a desktop tool that transcribes a video, uses a language model to correct and split the subtitle lines, then translates them and burns them into the video. It uses GPL-3.0. Choose it when you need subtitles and translation but not dubbing.
youtube-transcript-api is a Python library that fetches the transcript or auto-captions of a YouTube video. It is MIT licensed and handy for turning your own videos into blog posts, descriptions or clip ideas.
Upload and schedule your videos
Many YouTube automation tools stop at a finished video file, so you may need a separate tool for the last step.
youtubeuploader is a command-line YouTube auto uploader under Apache-2.0. It uploads a video with its title, description, thumbnail and scheduled publish time, which makes it easy to drop into a script. Upload tools that use the YouTube Data API need your own Google Cloud project and sign-in credentials. The tool’s README also warns that videos uploaded through API projects created after July 2020 stay private until Google audits the project.
social-auto-upload uploads videos to YouTube, TikTok, Douyin, Bilibili and other platforms on a schedule under MIT. Its project description is written in Chinese, so expect some translation work.
Postiz is a self-hosted social media scheduler with a calendar, AI captions, team approvals and analytics across many networks. It is AGPL-3.0 licensed and suits teams that post to several platforms, not only YouTube.
A plain note on YouTube’s policies
YouTube’s policies discourage mass-produced and repetitive videos, and its rules on monetisation and spam apply whether a video was made by hand or by a script. Tools can speed up production, but they cannot guarantee that a channel will be accepted for monetisation or stay in good standing. Add your own research, voice and editing, check the licences of any stock footage, music and AI voices you use, and read YouTube’s current policies yourself before you publish.
How to choose YouTube automation tools
Pick by the step that takes you the most time today:
- Producing new faceless videos: MoneyPrinterTurbo, or MoneyPrinter if you want to customise the code.
- Running a whole channel with review steps: youtube-automation-agent.
- Clipping long videos into Shorts: OpenShorts or AI-Youtube-Shorts-Generator.
- Reaching new languages: VideoLingo, or VideoCaptioner if you only need subtitles.
- Scheduled uploads: youtubeuploader for one channel, Postiz for many networks.
If you searched for YouTube automation on GitHub, the explorer button above lists every YouTube repo in our catalogue, and you can add workflow filters such as captions or auto-upload to narrow it.