Summary AutoShorts is a local-first desktop application that automates the creation of high-impact, vertical short-form video clips (9:16) from long-form content. It employs AI for viral moment detection and ranking, targeting content creators and editors seeking to efficiently repurpose media into engaging social content. The core benefit is a streamlined, automated workflow that significantly reduces manual clip generation effort.

How It Works Built with Tauri 2, React/TSX, Rust, and SQLite, AutoShorts offers a local-first approach. Its automated pipeline imports media, extracts audio, transcribes (Deepgram/local Whisper), and analyzes transcripts using dynamic LLM support (DeepSeek default, Claude option) to identify and rank viral moments. Native ffmpeg integration handles landscape-to-portrait auto-cropping for H.264 output.

Quick Start & Requirements

• Prerequisites: FFmpeg and FFprobe must be installed and on the system PATH. Instructions provided for macOS, Windows, and Linux.
• Installation: Download OS-specific packages (.dmg, .msi, .deb, .AppImage) from GitHub Releases and follow platform instructions.
• First-Launch: Onboarding wizard for offline (Ollama + Whisper) or cloud API key (Deepgram, DeepSeek, Claude) configuration.
• Development: npm install, npm run tauri:dev. Build: npm run tauri:build.
• Links: No direct documentation or demo links provided.

Highlighted Details

• Dynamic LLM Support: Integrates DeepSeek (default, cost-effective) and Claude (premium copywriting) for AI analysis.
• Automated Pipeline: End-to-end workflow from media import to ranked clip candidates.
• Local Data Storage: SQLite for local persistence of transcripts, candidates, and project data.
• Native Project Management: Built-in dashboard for managing video projects.
• AI-Powered Ranking: Identifies and ranks potential viral moments.
• Automated Cropping: Converts landscape video to 9:16 portrait using ffmpeg.

Maintenance & Community The provided README lacks specific details on maintainers, community channels (e.g., Discord, Slack), or project roadmap.

Licensing & Compatibility The README does not specify a software license, preventing an assessment of commercial use or closed-source integration compatibility.

Limitations & Caveats Using smaller local LLMs (e.g., 3B/7B Ollama) for viral moment detection is discouraged due to limited reasoning, potentially causing inaccurate timestamping and fragmented clips. Users may need to bypass OS security prompts (Gatekeeper, SmartScreen) for unsigned builds. The absence of a stated software license is a significant adoption blocker.