Agions/splicr

What it solves

splicr simplifies the creation of AI-driven video narrations and commentaries. It specifically targets creators making short dramas, movie recaps, and storytelling content for social media by automating the tedious process of cutting footage, writing scripts, and syncing audio.

How it works

The application uses a 7-step intelligent pipeline to transform raw footage into a finished video:

1. Import: Extracts metadata and generates thumbnails from local files.
2. Smart Cutting: Uses FFmpeg for scene detection and keyframe indexing.
3. AI Scripting: Generates first-person monologues or reviews using one of 11 supported LLMs.
4. TTS & Cloning: Converts scripts to speech using Edge-TTS, OpenAI-TTS, or GPT-SoVITS for voice cloning.
5. Subtitle Alignment: Employs VAD (Voice Activity Detection) via FFmpeg to align subtitles precisely.
6. Mixing: Aligns audio and video tracks with millisecond precision and manages background music levels.
7. Export: Produces final 1080P vertical videos or native CapCut draft files (.draft) for further editing.

Who it’s for

It is designed for self-media creators, short-drama editors, and movie commentators who want to speed up their production workflow while maintaining high viewer retention.

Highlights

• Broad LLM Support: Integrates 11 providers including Qwen, DeepSeek, GPT, Claude, Gemini, and local models via Ollama/LMStudio.
• Voice Cloning: Supports zero-shot voice cloning via GPT-SoVITS.
• CapCut Integration: Exports directly to CapCut draft format for professional post-production.
• High-Performance Stack: Built with Rust, Tauri 2, and React 19 for a responsive desktop experience.