jamesob
Featured
local-llm
A comprehensive guide for building and configuring a high-end local machine to run state-of-the-art LLMs, with detailed hardware choices, BIOS tuning, and Docker-based model serving.
AI repository intelligence
FR-AI / ANALYSISWhy it is worth attention
It integrates scene detection, vision AI, transcription, LLM script writing, local voice cloning, and an interactive multi-track timeline editor into a single tool, all reverse-engineered from a PyInstaller distribution.
Who it is for
- Video content creators and editors
- AI/ML developers experimenting with video processing
- Researchers in automated media production
- Developers needing a desktop video review pipeline
Use cases
- Automatically generating review scripts for movies or videos
- Transcribing and subtitling videos with SRT export
- Creating voiceovers using multiple TTS engines including voice cloning
- Editing multi-track video timelines with integrated audio and subtitles
Strengths
- Comprehensive all-in-one feature set covering scene analysis, transcription, script writing, and TTS
- Support for multiple LLMs (DeepSeek, OpenAI GPT, Google Gemini) and TTS engines (local and cloud)
- Local voice cloning capabilities via VieNeu and XTTS v2/GPT-SoVITS
- Interactive PyQt6 timeline editor with real-time video preview
Considerations
- Reverse-engineered code may contain errors that require self-fixing
- Windows-only with Python 3.9 recommended; GPU optional but beneficial
- Requires bundled FFmpeg binaries and may have large resource usage due to AI models
README quick start
AutoReviewTool - AI Video Review & Automated Subtitle/TTS Studio
The source code has been analyzed and reverse-engineered—if you encounter errors, fix them yourself.
📌 Overview
AutoReviewTool is a comprehensive PyQt6 desktop application designed for automated movie/video review generation, video re-captioning, and AI voice synthesis. It combines computer vision, speech recognition, large language models (LLMs), multi-engine text-to-speech (TTS), and an interactive multi-track video timeline editor into a unified studio environment.
The codebase was decompiled and reverse-engineered from a PyInstaller distribution, and subsequently refactored for stability, clean execution, and zero-crash operation under Python 3.9+.
🚀 Key Features
-
🎬 Automated Video & Scene Analysis
- Scene boundary detection using
PySceneDetect. - Vision AI frame analysis via Google Gemini & OpenAI GPT-4 Vision.
- Automatic video chaptering and plot beat extraction.
- Scene boundary detection using
-
🗣️ Audio Transcription & Subtitling
- Speech-to-text transcription powered by
Faster-Whisperandopenai-whisper. - Subtitle segmentation, re-chunking (
_rechunk_srt), and SRT export. - Intelligent multi-language support with automatic language detection.
- Speech-to-text transcription powered by
-
📝 AI Script Writing & Review Generation
- Multi-LLM integration supporting DeepSeek, OpenAI GPT, and Google Gemini APIs.
- Automatic review script writing tailored to specific target durations (e.g., 3-15 minutes).
- Interactive script editing interface (
SubtitleScriptPanel) with instant voice preview.
-
🎙️ Multi-Engine Text-to-Speech (TTS) & Voice Cloning
- Local Offline Voice Cloning via VieNeu (
vieneu_worker.py) and XTTS v2 / GPT-SoVITS (xtts_worker.py). - Cloud TTS engines including Edge-TTS, FPT AI, Viettel, gTTS, and ElevenLabs.
- Per-scene TTS generation with automatic fallback handling and audio duration matching.
- Local Offline Voice Cloning via VieNeu (
-
🎛️ Interactive Multi-Track Timeline Editor
- Custom PyQt6 painter canvas (
_TLCanvas) for visual timeline editing. - Multi-track layout: Video Track, Subtitle/Voiceover Track, Background Music Track, SFX Track.
- Real-time video preview (
VideoPreview) synchronized with timeline seeking and playback.
- Custom PyQt6 painter canvas (
-
💾 Session Persistence & Exporting
- Automatic session autosave and recovery (`session_
Description
AUTOREVIEWTOOL (The source code has been analyzed and reverse-engineered—if you encounter errors, fix them yourself.)
Related repositories
Similar projects matched by category, topics, and programming language.
slvDev
Featured
esp32-ai
A 28.9 million parameter language model runs on an $8 ESP32-S3 microcontroller entirely on-device, generating simple stories at about 9.5 tokens per second.
gavamedia
Featured
deltafin
Deltafin is a research project that runs the 2.8-trillion-parameter Mixture-of-Experts model Kimi K3 on a single Apple Silicon Mac (e.g., M1 Max with 64 GB) at about 16 seconds per token, using exact, reproducible inference with local or streaming expert loading.