OpenShorts is an open-source, MIT-licensed AI video clipping platform that converts long-form YouTube videos, podcasts, and livestreams into vertical 9:16 shorts for TikTok, Instagram Reels, and YouTube Shorts. Creators can paste a YouTube link to generate captioned vertical clips in minutes, run the cloud web app, or self-host the entire processing pipeline for free using Docker.
By combining Google Gemini for transcript reasoning with faster-whisper for speech recognition, OpenShorts eliminates the subscription lock-in common to commercial clipping services.
Related Articles: Opus Clip AI Clipper | Fliki AI Video Creator | Autodraft AI Animation | D-ID AI Avatars | Vidu AI | Kling AI
Key Takeaways: OpenShorts AI Clip Generator
- OpenShorts is licensed under MIT, enabling free and unlimited self-hosting on local machines or servers using Docker and Docker Compose.
- The platform analyzes long videos with Google Gemini or local LLMs to score viral potential out of 100 and extract 3 to 15 clips between 15 and 60 seconds.
- Smart 9:16 reframing includes automatic face tracking and multiple framing layouts like Track, Split, and Screencast.
- Integrated faster-whisper produces word-level dynamic subtitles, with ElevenLabs support for multi-language voice dubbing.
- A Model Context Protocol (MCP) server and REST API allow autonomous AI agents to automate end-to-end video clipping and social publishing.
What is OpenShorts?
OpenShorts is a video automation system designed to repurpose horizontal long-form footage into short-form vertical assets. Commercial clipping tools often impose strict monthly minute limits, expensive tiers, and recurring watermark restrictions. OpenShorts solves this by publishing its core codebase under the permissive MIT license on GitHub (https://github.com/mutonby/openshorts) while also providing a hosted web app at https://www.openshorts.app/.
The application handles the entire short-form pipeline: audio transcription, natural language context analysis, viral moment detection, speaker-focused vertical framing, subtitle burning, and social publishing.
Core features of OpenShorts
OpenShorts incorporates several specialized media processing modules to automate short creation:
- AI moment detection: Scans transcripts using large language models like Google Gemini to detect engaging hooks, jokes, and key takeaways. The model rates segments on a 1-to-100 viral score and isolates 3 to 15 clips lasting 15 to 60 seconds.
- Smart 9:16 reframing: Uses computer vision face tracking to keep active speakers centered. Layout modes include Single Track, General, Two-Speaker Split (for interview podcasts), and Screencast (for tutorials).
- Dynamic subtitles and dubbing: Employs faster-whisper to generate synchronized, word-by-word animated subtitles. It also integrates ElevenLabs for synthetic voiceovers and voice dubbing across multiple languages.
- AI Shorts Creator: Generates user-generated content (UGC) style promo videos with synthetic AI actors and voices through fal.ai integrations.
- YouTube Studio assistant: Drafts search-friendly video titles, descriptions, timestamps, and thumbnail concepts to accompany each cut.
- Auto-publishing and agent workflows: Supports direct social posting via webhooks, REST APIs, and services like Upload-Post. Its native MCP server lets developer agents trigger clipping tasks programmatically.
How to generate vertical clips using OpenShorts
Transforming a long video into vertical shorts takes five straightforward steps:
Submit video source
Paste a public YouTube link or upload a local MP4 file into the OpenShorts web interface or your local instance.
Run AI moment detection
The engine downloads the audio track, runs faster-whisper transcription, and passes the text to Gemini to detect high-interest sections.
Choose reframing layout
Select your 9:16 layout preset: Single Track for solitary speakers, Two-Speaker Split for interview dialogue, or Screencast for software demonstrations.
Customize subtitles and audio
Review the generated word-by-word captions, adjust font styles and highlight colors, and configure optional ElevenLabs dubbing.
Export and distribute
Render the vertical clips in 1080p resolution, download them locally, or push them to social channels using automated webhooks.
Self-Hosting with Docker for Zero Watermarks and Unlimited Minutes
Running OpenShorts through Docker Compose gives you unlimited rendering with zero platform fees and no watermarks. You only supply your own Google Gemini API key, which includes generous free monthly request quotas.
Hosting options: self-hosted Docker vs cloud app
OpenShorts provides two distinct operating models to fit different creator technical requirements:
| Dimension | Self-Hosted (Docker) | Cloud Free Tier | Cloud Paid Plans |
|---|---|---|---|
| Monthly Cost | $0 (Open Source MIT) | $0 / month | From $12 to $59 / month |
| Watermark | None | Yes (OpenShorts badge) | None |
| Clip Storage | Local storage / Unlimited | 7-day expiration | Extended cloud storage |
| Video Minute Limit | Unlimited | Limited trial minutes | High-capacity allowances |
| Hardware Required | Local PC or VPS (Docker) | None (Browser only) | None (Browser only) |
| Rendering Speed | Depends on local CPU/GPU | Standard queue | High-speed dedicated cloud GPUs |
| API Keys | Bring your own Gemini key | Included | Included |
Supported use cases and content genres
OpenShorts handles diverse video formats across several production scenarios:
- Podcast clipping: Detects dialogue exchanges between hosts and guests, using Split layout to show both speakers simultaneously in vertical 9:16.
- Gaming highlights: Isolates high-energy gameplay moments, kills, and reactions from GTA 5, GTA 6, World of Warcraft raids, and console capture sessions.
- Twitch and Kick livestreams: Scans hours-long stream recordings to pull funny moments and highlights for TikTok and Instagram Reels.
- Sermons and speeches: Extracts memorable quotes and moral takeaways from church sermons and keynote presentations.
- Educational tutorials: Uses the Screencast reframing mode to keep both code demonstrations and instructor reactions legible on mobile screens.
- AI UGC video advertisements: Combines synthetic voices with fal.ai video avatars to produce short social video ads without camera setups.
OpenShorts vs leading alternatives compared
OpenShorts differs from proprietary clipping platforms like Opus Clip primarily through its open-source license and local Docker execution:
| Feature / Tool | OpenShorts | Opus Clip | Klap | Vizard | Submagic |
|---|---|---|---|---|---|
| License | Open Source (MIT) | Proprietary | Proprietary | Proprietary | Proprietary |
| Self-Hosting | Yes (Docker) | No | No | No | No |
| Starting Price | Free / $12 mo | $9 / mo | $29 / mo | $16 / mo | $20 / mo |
| Moment Detection | Google Gemini / Local | OpusClip 4.0 AI | Proprietary AI | Proprietary AI | Proprietary AI |
| Face Tracking 9:16 | Yes (Track & Split) | Yes (Track & Split) | Yes | Yes | Yes |
| Word-Level Captions | Yes (faster-whisper) | Yes (Animated) | Yes | Yes | Yes |
| AI Agent Support | MCP Server & REST API | REST API | No API | Limited API | Limited API |
OpenShorts Advantages
5 items- Completely free and unlimited when self-hosted under the MIT license.
- Eliminates subscription lock-in and vendor pricing hikes.
- Integrated Google Gemini reasoning identifies logical narrative hooks accurately.
- Offers multiple vertical layout options, including two-speaker split screens.
- Includes an MCP server and developer API for AI agent automation.
OpenShorts Considerations
3 items- Self-hosting requires basic familiarity with Docker, environment variables, and API keys.
- Rendering speed on local hardware depends heavily on your system processor and GPU capabilities.
- Cloud free tier includes watermarks and a 7-day clip expiration window.
