VideoCaptioner is an LLM-powered video subtitle processing tool that provides an all-in-one workflow for speech recognition, subtitle optimization, translation, and video synthesis. It can transcribe audio and video into subtitles, improve subtitle segmentation and accuracy, translate subtitles into other languages, and synthesize the processed subtitles back into videos.

The tool supports multiple speech recognition engines and translation services, including free options that require no API key. LLM functionality can be used for semantic subtitle optimization and context-aware translation, while configuration can be managed through command-line arguments, environment variables, configuration files, or default settings.

VideoCaptioner uses word-level timestamps and VAD (Voice Activity Detection) to improve speech recognition accuracy. It also uses LLM semantic understanding for more natural subtitle segmentation, supports context-aware translation with reflection-based optimization, and provides concurrent batch processing for efficient video and subtitle workflows.