JayWebtech/autoshorts

What it solves

AutoShorts is a desktop application designed to automate the process of turning long-form video or audio recordings into short-form, vertical clips (9:16) optimized for social media. It eliminates the manual effort of scanning long recordings to find viral-worthy moments and manually cropping landscape video to portrait.

How it works

The app uses an automated pipeline that imports media, extracts audio, and transcribes the content using Deepgram or local Whisper. It then leverages LLMs (such as DeepSeek, Claude, or local models via Ollama) to analyze the transcript, detect viral moments, and rank them. Finally, it uses native FFmpeg integration to automatically center-crop landscape videos into vertical H.264 portrait clips.

Who it’s for

Content creators and podcasters who need to efficiently convert long-form content into high-impact short-form clips for platforms like TikTok, Reels, or Shorts.

Highlights

• Multi-LLM Support: Flexible choice between DeepSeek, Claude, and local Ollama models for moment detection and hook analysis.
• Automated Pipeline: Single-chain workflow from media import and transcription to AI analysis and ranking.
• Local-First Design: Uses a local SQLite database to store transcripts, candidates, and project data.
• Portrait Auto-Cropping: Native FFmpeg integration for automatic landscape-to-vertical cropping.
• Flexible Deployment: Supports both cloud API-based workflows and fully offline local setups.