BibiGPT

Overview

BibiGPT is an AI-powered audio/video summarizer and knowledge assistant that turns any video, podcast, meeting recording, or local media file into structured, searchable, and reusable knowledge artifacts. Originally launched as BiliGPT in 2023 by JimmyLv (吕立青), the platform has served over one million users worldwide. By simply pasting a link or uploading a file, users can obtain a time-stamped summary, a full transcript, a mind map, and an interactive AI chat interface grounded in the source content — transforming passive viewing into an active, efficient learning experience.

At its core, BibiGPT leverages large language models to automatically extract transcripts and generate structured, chapter-based summaries with precise timestamps, enabling users to jump directly to the most relevant sections. The platform supports multi-turn AI chat grounded in the video content, allowing users to ask follow-up questions and clarify complex concepts. Advanced capabilities include visual content analysis for frame-by-frame understanding of slide-heavy presentations, subtitle translation across all language permutations (Chinese, English, Japanese, Korean, and Traditional Chinese), subtitle burn-in, and article repurposing that rewrites video content into posts optimized for WeChat, Xiaohongshu (RED), Twitter/X, and blogs.

BibiGPT supports more than 30 video and audio platforms, including YouTube, Bilibili, TikTok, Douyin, Kuaishou, X/Twitter, Vimeo, TED, Coursera, Apple Podcasts, Spotify, and Khan Academy, as well as arbitrary web pages with embedded audio/video tags and local files such as mp3, mp4, m4a, wav, srt, vtt, and ass. The service is available on Web, browser extensions (Chrome, Firefox, Edge), desktop applications (macOS and Windows), and mobile apps (iOS and Android). For developers and AI agents, BibiGPT exposes an MCP server with OAuth 2.0 authentication, a REST OpenAPI specification, and an NLWeb /ask endpoint with SSE streaming, enabling seamless integration into automated knowledge pipelines.

BibiGPT differentiates itself with native support for Chinese platforms, first-class Agent/MCP integration, and robust cross-platform coverage. It offers a free tier with a daily quota and a 30-minute video length cap, while paid plans (Plus, Pro, and Lifetime) unlock advanced features such as batch processing, visual content analysis, and full API access. Regarding privacy, uploaded files are automatically deleted after 30 days, summaries are private by default, and user content is not used for training without explicit opt-in. With multi-language support, one-click export to Notion, Obsidian, Feishu, and flomo, and a focus on turning knowledge into actionable output, BibiGPT is designed for learners, researchers, content creators, and developers who need to process audio and video content at scale.

Core capabilities

Timestamped Video Summarization

Generates structured, chapter-based summaries with precise timestamps so users can navigate long videos and jump straight to key sections.

Full Transcript Extraction

Extracts complete transcripts and subtitles from videos, podcasts, and local files across 30+ platforms, supporting mp3, mp4, m4a, wav, srt, vtt, and ass formats.

Mind Map Generation

Automatically converts video and audio content into structured mind maps for visual knowledge organization and rapid review.

Grounded AI Chat

Provides multi-turn question-answering grounded in the source video content, enabling users to ask follow-up questions and gain deeper understanding.

Visual Content Analysis

Performs frame-by-frame understanding of slides and on-screen visuals, making it ideal for presentation-heavy and slide-based videos.

Article Repurposing

Rewrites video content into articles optimized for WeChat, Xiaohongshu (RED), Twitter/X, and blogs for multi-platform publishing.

Subtitle Translation & Burn-in

Supports translation across all language permutations (Chinese, English, Japanese, Korean, Traditional Chinese) and burns translated subtitles directly into video.

Knowledge Base Sync

One-click export and push to Notion, Obsidian, Feishu, and flomo, preserving Markdown formatting, timestamps, chapters, and source links.

Agent & API Integration

Offers an MCP server with OAuth 2.0, a REST OpenAPI endpoint, and an NLWeb /ask endpoint with SSE streaming for AI agent and developer integration.

Multi-Platform Accessibility

Available on Web, browser extensions (Chrome/Firefox/Edge), desktop (macOS/Windows), and mobile (iOS/Android).

Use cases

Summarizing long YouTube tutorials or Bilibili lectures into timestamped chapter summaries for rapid review.

Converting podcast episodes into searchable transcripts, mind maps, and structured notes for knowledge management.

Analyzing slide-heavy conference talks and presentations through frame-by-frame visual understanding.

Repurposing video content into blog posts or social media articles for multi-platform distribution.

Translating and burning subtitles for cross-language viewing and global content distribution.

Syncing AI-generated notes and summaries to Notion, Obsidian, or Feishu for personal knowledge bases.

Building AI agents that call video summarization tools via MCP server or REST API for automated content pipelines.