Bilive is a Python tool for automated Bilibili live stream recording, processing, and uploading that handles capture, danmaku rendering, subtitle generation, video slicing, and submission to the platform.
The tool solves the problem of manually monitoring and recording live streams by automating the entire workflow from capture through publication. It listens for Bilibili live broadcasts, records both video and danmaku comments including paid messages and gifts, then processes the recordings by converting danmaku to match resolution, performing speech recognition for subtitles, and rendering these elements into the video. It automatically identifies highlight segments based on danmaku density, uses multimodal language models to generate titles and descriptions for clips, and generates video thumbnails via image generation models before uploading finished content back to Bilibili. The tool is designed to run unattended around the clock and supports low-resource deployments including machines without GPUs and both x64 and ARM64 architectures.
Bilive suits anyone who wants to automate live stream archival and clip generation for Bilibili without manual intervention or expensive hardware. It works well for content creators who want to repurpose live broadcasts into edited highlight videos and for archivists managing multiple streams. The tool integrates with multiple large language model providers including OpenAI Whisper, Zhipu GLM-4V-PLUS, Google Gemini 1.5 Pro, Qwen-2.5-72B-Instruct, and SenseNova V6 Pro, allowing flexibility in model selection based on availability and cost.
The project shows active development with regular updates addressing both core functionality and compatibility across different hardware configurations. Work spans the full pipeline from stream detection through final upload, indicating sustained effort across multiple system components. The codebase maintains support for resource-constrained environments alongside standard deployments, suggesting ongoing attention to accessibility and practical deployment scenarios.
