social-media-copilot (社媒助手开源版) is a GPL-3.0 TypeScript browser extension that collects data from Xiaohongshu (小红书), Douyin (抖音) and Kuaishou (快手), built for developers, researchers and small operators who need self-hosted social media data collection instead of paying for a hosted product or scripting it by hand.

What it is

The project is a Chrome and Edge browser extension, written in TypeScript and built on the WXT framework, that runs inside the user's own browser session on the three supported platforms. It ships as three editions from one repository: the closed-source 商店版 with the widest platform coverage, the 开源版 on the main branch, and the server branch, which wraps the extension in an HTTP service for programmatic use. The registry entry tracks the main branch, described in the README as derived from the v0.x codebase of 社媒助手 and maintained for basic functionality and learning exchange.

The concrete problem it solves is the cost and friction of routine collection work. The README states the situation plainly: at low volumes, manual handling is tedious, while other products may be expensive or demand a complicated development environment to configure. The extension replaces both paths. Installing it in a browser is enough to reach a self-sufficient collection workflow, with no extra setup, and the server branch adds an HTTP surface for callers who need data by API rather than by clicking through a popup. The repository has 1,329 stars and 213 forks, with two open issues at the recorded last push of 2026-08-31.

Key capabilities

• Watermark-free video and image download on 小红书, 抖音 and 快手.
• Collection of data for a specified work (采集指定作品的数据) on all three platforms.
• Collection of comments for a specified work (采集指定作品的评论).
• Collection of data for a specified creator (采集指定创作者的数据).
• Collection of a specified creator's works (采集指定创作者的作品).
• Per-platform extension entrypoints — xhs.content, dy.content and ks.content — alongside background and popup, under src/entrypoints/.
• The server branch exposes an HTTP service that proxies requests to the extension side over socket.io; the extension issues the real request and calls the result back to the server.

Who uses it and how

• Operators whose collection volume is small enough that manual copying is wasteful but who do not want to buy a hosted scraping product, and who therefore install the extension and work from the browser directly.
• Teams that need programmatic access deploy the server branch, connect the extension to that server, and then call the server-side HTTP API to obtain platform data.
• Developers studying browser extension architecture build the main branch locally and load the unpacked output from output/chrome-mv3, using the code as reference material rather than as a production dependency.
• Local development runs against a NodeJS toolchain with pnpm dev, with documentation per platform in docs/dy.md, docs/xhs.md and docs/ks.md.

Getting started

NodeJS is required, with pnpm recommended although npm, yarn and cnpm also work; run pnpm install then pnpm dev, open chrome://extensions/, enable developer mode, and load the unpacked extension from the output/chrome-mv3 directory. The 开源版 is also published on the Chrome Web Store, and after the extension connects to the server side, platform data can be fetched through the server API.

How it compares

No paid product is named anywhere in the supplied facts, and no comparable third-party tool is named either, so this project stands alone in this registry. The only alternatives the README does name are its own siblings: the closed-source 商店版, which covers 小红书, 抖音, 哔哩哔哩, 快手 and TikTok and is maintained first, and the server branch for API callers. The open-source edition is positioned as the learning-oriented one, not as the most complete.

When to use it — and when not to

A self-hoster must operate the NodeJS and pnpm build chain, build the extension, and keep a browser running with the extension connected to the server, because the server only proxies and the extension performs the actual requests. Commercial use should be ruled out: the disclaimer restricts the project to study and research, forbids commercial activity, and asks that it be deleted within 24 hours of download, while also warning of risks including device damage and account bans. The Docker deployment mentioned in the tagline has no documented image, compose file or command in the README excerpt, and the open-source edition is tied to the older v0.x code with only basic maintenance promised, so anyone needing full platform coverage or ongoing fixes should look at the closed-source edition instead.

.
├── src/                    # 源代码
│   ├── assets/             # 资源目录
│   ├── components/         # 公共组件
│   ├── entrypoints/        # 插件入口
│   │   ├── background/     # 插件的background
│   │   ├── popup/          # 插件的popup
│   │   ├── xhs.content/    # 小红书平台相关脚本
│   │   └── dy.content/     # 抖音平台相关脚本
│   │   └── ks.content/     # 快手平台相关脚本
│   ├── public/             # 静态文件
│   └── utils/              # 工具函数
├── output/                 # 项目编译输出目录
├── docs/                   # 文档与截图
│   ├── images/             # 项目截图
│   ├── dy.md               # 抖音功能介绍
│   └── xhs.md              # 小红书功能介绍
│   ├── ks.md               # 快手功能介绍
├── package.json            # 项目依赖配置
├── README.md               # 项目说明
└── wxt.config.ts           # WXT配置文件

# 安装依赖
pnpm install
# 运行项目
pnpm dev