OpenShorts 🧩 Workflow Open source
Self-hostable clip generator, AI UGC video maker and YouTube toolkit with an MCP endpoint
- GitHub stars
- 5.6k
- Stars this week
- –
- Forks
- 1.3k
- Licence
- MIT
- Last push
- 2026-09-23
- Maintainer
- mutonby
git clone https://github.com/mutonby/openshorts.gitThird-party workflows & skills run with your permissions and API keys. Read the source before installing, and prefer pinned versions.
Works with
Built on the open Agent Skills standard, so it works in every app that supports Agent Skills.
Good for
About OpenShorts
What it is
OpenShorts by mutonby (MIT) is an open-source video platform with three tools: a Clip Generator that turns podcasts, webinars and vlogs into 9:16 shorts with face tracking and captions; AI Shorts, which creates marketing videos with AI actors, voiceover, b-roll and subtitles; and a YouTube Studio for thumbnails, titles and descriptions. The same software runs self-hosted or on the hosted openshorts.app service.
Key features
- Clip Generator with face-tracked crops and stacked two-speaker layouts
- Captions placed to avoid covering faces
- AI Shorts UGC-style videos from a product description or URL
- YouTube thumbnails, titles and descriptions
- Direct publishing to TikTok, Instagram and YouTube
- MCP endpoint and API for agents
How to install
Clone with git clone https://github.com/mutonby/openshorts.git, add Gemini, ElevenLabs and fal.ai keys, and run docker compose up --build. It needs 8 GB+ RAM; an NVIDIA GPU speeds processing.
Pricing and rights
Self-hosting is free under MIT with no watermark or limits; you pay for your own API usage (the README estimates about $0.65-$2 per AI Shorts video). The hosted service has a free plan with a watermark and 20 minutes a month, and paid plans from $12/month.
Who it is for
Podcasters and small businesses who want OpusClip-style clipping and simple video ads without a subscription.
Verdict
A capable free alternative to paid clipping tools, actively maintained. The AI-actor feature produces synthetic testimonials that should be clearly disclosed as AI, and self-hosting on CPU is slow.
Pros
- Free self-hosted clipping without watermark
- Face-aware layouts and captions
- MCP endpoint for agents
- MIT
Cons
- AI-actor ads need clear disclosure
- Slow on CPU
- Several API keys required
Similar workflows & skills
All open-source content pipelines →GPT Researcher 🧩 WorkflowOpen source
Autonomous research agent that writes cited, long-form reports from web and local sources
Open Notebook 🧩 WorkflowOpen source
Self-hosted NotebookLM alternative for research notes and multi-speaker podcasts
WhisperX 🧩 WorkflowOpen source
Fast Whisper transcription with word-level timestamps and speaker diarization
VideoLingo 🧩 WorkflowOpen source
Netflix-style subtitle cutting, translation, alignment and dubbing for videos
pyVideoTrans 🧩 WorkflowOpen source
Translate and dub videos: speech recognition, subtitle translation and TTS
Podcastfy 🧩 WorkflowOpen source
Python package that turns web pages, PDFs and videos into AI podcast conversations