yotsuki2213/bilibili_videoread_mcp

This page was translated automatically. Technical details may differ from the original. View the original

43 stars 0 forks Quality 78 Trend 26MIT

Bilibili Video Subtitles/Bullet Comments/Comment Extraction MCP Server — Login by scanning QR codes, automatically extracting subtitles/bullet comments/comments and metadata, and outputting structured Markdown for AI understanding

Overview

~~Let AI also watch Bilibili~~ Extract Bilibili videos from your AI with one click: subtitles, bullet comments, comments, and metadata, output structured Markdown for LLMs to digest. Your AI assistant (Claude, ChatGPT, Cherry Studio... ) No more staring blankly at Bilibili links. Just enter a BV account, and subtitles, bullet comments, titles, uploaders, and views will automatically appear. After reading it, AI will summarize, refine, translate, and recreate for you— ⚡ 🚫 : This project uses Bilibili's official QR code scanning login; the terminal scan the code and the phone automatically logs in; After about 30 days, the scan is re-scanned, maintenance-free throughout the process, and no need to restart the session. - 🎯 — Opencode Lightweight Subscription ($10/month): https://opencode.ai/go?ref=G9NS3SGZMT - ☁️ — High cost-performance stable airport: https://xn--9kqz23b19z.com/#/register?code=LBcBtWRW - Credentials exist **(Windows: %APPDATA%\videoread\auth.json, Unix: ~/.config/videoread/auth.json), will not be written to projects or .mcp.json - SESSDATA expires in about 30 days; after expiration, run login again and scan the code - Common command: python -m videoread_mcp status Check login status and remaining days, log out and log out 💡 It is recommended to use an absolute path to point to the project venv:C:...\BiliBili_VideoRead_MCP.venv\Scripts\python.exe Help me summarize what this video is about: [https://www.bilibili.com/video/BV1GJ411x7h7] And by the way, let's see what people are saying in the comments AI automatically adjusts extract_bilibili_video/extract_bilibili_comments, instantly generating content.

Overview (machine translated)

~~Let AI also watch Bilibili~~ Chew up Bilibili videos chew up to your AI: subtitles, bullet comments, comments, metadata extracted with one click, output structured Markdown for LLMs to digest.

Your AI assistant (Claude, ChatGPT, Cherry Studio… ) No more staring blankly at Bilibili links. Just enter with a BV account, and subtitles, bullet comments, titles, uploaders, and views will automatically appear. After the AI reads, it can summarize, extract, translate, and create secondary creations for you—Leave the coin insertion to humans, and the video watching to the AI ⚡

🚫 > Say goodbye to F12 handwritten cookies: This project uses Bilibili’s official QR code scanning login; the terminal scan and the phone automatically logs in; After about 30 days, re-scan is required, maintenance-free, no need to restart the session.

⭐ Please give a Star to support us—the author will be more motivated to update and maintain it!

🤝 Friendly Links / SUPPORT

🎯 OpenCode Go — OpenCode Lightweight Subscription ($10/month):
☁️ Conscience Cloud — High cost-performance stable airport:

By registering with an invitation code, authors can earn rebates and be more motivated 🫡 to maintain and update their work

✨ Feature Introduction

Features	Introduction
🔗 Link Recognition	BV / AV number, full link, short link `b23.tv` all accessible
📋 Metadata	Title, UP creator, duration, views, likes/coins/favorites, release date, description, cover
📑 Multi-P	Metadata contains the entire P list, allowing extraction of all or specified P subtitles
💬 Subtitles	Prioritize Chinese AI subtitles, automatically revert language; Output by semantic segmentation
🎯 Danmaku	Full Danmaku (segmented interface, non-real-time pool), sorted by time, automatic over-limit uniform sampling
💭 Comments	Sorted by popularity / time, optional with sub-replies
📄 Markdown	All output is structured, and AI can instantly start chatting at a glance

🚀 Get Started Quickly

1. Installation

2. Scan the QR code to log in (one-time, completed in 10 seconds)

The terminal will print a QR code. Open the Bilibili app on your phone → scan the top right corner, and after a few seconds, the login status will automatically be saved:

The credential is stored in your local user directory (Windows: %APPDATA%\videoread\auth.json, Unix: ~/.config/videoread/auth.json), and will not be written to the project or .mcp.json

SESSDATA expires in about 30 days. After it expires, run again and scan the code login.

Common commands: python -m videoread_mcp status Check login status and remaining days, logout Log out

3. Connect to the MCP client

Take Claude / Cherry Studio as an example: No environment variables required:

This page was translated automatically. Technical details may differ from the original.

Install

This server does not publish a one-line install command.

Open the repository installation guide New to MCP? Start with the mcpskill.org guide →

Configuration

{
  "mcpServers": {
    "videoread": {
      "command": "python",
      "args": ["-m", "videoread_mcp"]
    }
  }
}