lycohana/BiliSum

为 Bilibili、YouTube 及本地视频提供 AI 视频摘要和知识库.AI video summarizer and knowledge base for Bilibili, YouTube and local videos.

What it solves

BiliSum addresses the difficulty of extracting and organizing knowledge from long-form video content. It automates the process of converting videos from Bilibili, YouTube, or local files into structured text, visual notes, and searchable knowledge bases, eliminating the need for manual transcription and note-taking.

How it works

The tool follows a pipeline: it imports video content, uses ASR (Automatic Speech Recognition) for transcription, and then employs LLMs and VLMs (Vision Language Models) to generate summaries, structured notes, and mind maps. It can extract key frames to create "VLM-understood" visual notes where images are integrated into the text based on visual context. For long-term storage, it uses embeddings to create a local RAG (Retrieval-Augmented Generation) system for semantic search and Q&A across multiple videos.

Who it’s for

It is designed for students, researchers, and lifelong learners who consume educational or informational video content and want to transform that content into a permanent, searchable personal knowledge base.

Highlights

  • Multimodal Note-taking: Generates visual notes by combining transcriptions with AI-selected video screenshots.
  • Local-First Privacy: Data, indices, and configurations are stored locally, with support for local LLMs and embeddings.
  • Flexible Input: Supports Bilibili, YouTube, and local video formats (mp4, mkv, mov, webm).
  • Agent Integration: Provides a CLI and a specific "skill" that allows AI agents (like Claude or Cursor) to trigger summaries and transcriptions.
  • Knowledge Management: Features a RAG-based knowledge base for cross-video semantic retrieval and mind map generation.

Related

  • Project
    JefferyHcool/BiliNote

    An AI-powered video note-taking tool that converts videos from Bilibili, YouTube, and other platforms into structured Markdown notes with automatic summaries and screenshots.

  • Project
    via007/bilibili-rag

    A RAG-based system that converts Bilibili video collections into a conversational knowledge base by transcribing audio to text and indexing it for semantic search.

  • Project
    zrt-ai-lab/ViNote

    An AI-powered video-to-knowledge tool that converts videos from YouTube, Bilibili, or local files into structured notes, mind maps, and searchable knowledge assets.

  • Project
    wendy7756/AI-Video-Transcriber

    AI Video Transcriber is a FastAPI‑based web app (with CLI/MCP support) that extracts subtitles or runs Whisper on YouTube, TikTok, Bilibili, podcasts, etc., then uses an OpenAI‑compatible LLM to clean up the transcript, translate it if needed, and generate a multilingual summary. It can also download the original video in parallel, works via Docker or a simple install script, and offers a REST API for programmatic use.

  • Project
    Anil-matcha/AI-Youtube-Shorts-Generator

    AI‑YouTube‑Shorts‑Generator is an open‑source Python tool that turns long YouTube videos (or local files) into vertical short clips ready for TikTok/Reels/YouTube Shorts. It downloads the video, transcribes it with Whisper, uses an LLM (OpenAI, Gemini, or MuAPI’s model) to rank “viral‑worthy” moments, deduplicates overlapping candidates, and auto‑crops each highlight to 9:16 (or any aspect ratio). Two modes are offered: a zero‑setup API mode that delegates download, transcription, ranking and cropping to MuAPI, and a fully local mode that runs `yt‑dlp`, `faster‑whisper`, OpenCV/ffmpeg and calls an external LLM. The CLI returns MP4 files plus a score, hook sentence and reasoning for each clip, and can also output a detailed JSON file. MIT‑licensed, suitable for creators, agencies, or developers building their own clipping SaaS.