Open-source project
3441293738/creatorhub avatar
3441293738/creatorhub

CreatorHub puts four Chinese short-video platforms behind one local FastAPI panel

多平台内容监控·采集·搬运 一个 Web 面板管起抖音 / 小红书 / 快手

2,077 stars381 forksPythonLicense varies

At a glance

What is it?
3441293738/creatorhub is a self-hosted Python and FastAPI web panel for managing Douyin, Xiaohongshu, Kuaishou and WeChat Channels accounts: monitoring posts and comments, bulk keyword collection, downloading media, publishing and forwarding, and pushing alerts. Platform support is uneven, and the repository ships with no licence.
Who is it for?
Adopt it if you run several accounts across these platforms and want one dashboard for watching posts and comments, pulling media and republishing it elsewhere, with everything stored on your own machine. Do not adopt it if you need a licence grant, since there is no LICENSE file in the tree and GitHub reports no licence, or if your workflow depends on Xiaohongshu keyword collection, which is still listed as planned.
Can I use it commercially?
Not without permission. GitHub finds no licence file in the repository, and without a licence all rights are reserved by default: you may read the code but not reuse it. Check the README, or ask the authors, before using it.
Is it still maintained?
Yes. The repository last received commits 1 day ago.
What is it written in?
Mainly Python, according to GitHub's language statistics.

Answers come from the project's GitHub data, last synced on September 17, 2026, and from our analysis. They are not legal advice.

DEEP OPEN-SOURCE ANALYSIS

A local panel, not a hosted service

CreatorHub is a web application you run yourself. Python and FastAPI serve a single interface for managing accounts, watching posts and comments, downloading content, publishing, and receiving notifications, and the README is explicit that login state, the database and downloaded media all stay on your machine.

Four platforms are covered: Douyin, Xiaohongshu, Kuaishou and WeChat Channels. There is a public preview hosted on GitHub Pages, but the README states it uses desensitised sample data and only demonstrates the interface, since logging in, scraping, downloading and publishing all require running locally.

The audience is a creator or a small agency operating several accounts at once. The value proposition is consolidation rather than any single capability: one place to see what happened across four platforms instead of four apps.

The capability matrix is uneven, and that is the point

The README's own feature table is the most useful part of the documentation, because it is honest about what does not work. WeChat Channels is the clearest example: login by QR code, publishing, and management of your own works, data and comments, while monitoring and downloading anyone else's content is not supported at all. The note attached to it says Channels only covers your own account data inside the creator assistant.

Douyin is the deepest integration. It has keyword bulk collection, post and comment monitoring, danmaku monitoring on both the playback page and the creator centre, content download with selectable quality, publishing, automatic commenting and replying, and full own-account management across works, follows, fans and direct messages.

Xiaohongshu and Kuaishou sit in between: login, post and comment monitoring, download, publishing, and automatic comment and reply, with Xiaohongshu monitoring covering creator pages and keywords. Keyword bulk collection is Douyin only for now, and the README lists the Xiaohongshu version as planned.

Notifications go out through Bark, DingTalk or Telegram on all four platforms.

One browser profile per account, and why

Browser interaction runs either through the system Chrome over the Chrome DevTools Protocol or through Patchright, which the README describes as free and open source, with the business layer written against the Playwright API. Each account gets its own persistent browser profile so cookies, cache and local storage are never shared between accounts.

That isolation is the design decision worth understanding, because it is what makes multi-account operation viable at all. Sharing one profile would cross-contaminate login state and make platform-side rate limiting unpredictable. The README recommends the system's stable Chrome for Xiaohongshu QR login and says it will fall back to a visible Patchright Chromium when Chrome is not installed.

Per-account proxy configuration is exposed in the interface as a separate pool associated with individual accounts, which fits the same isolation model.

It is also worth being clear-eyed about what this means operationally: this is browser automation against consumer platforms, and the accounts you drive are the accounts that carry the consequences.

Getting it running

On Windows x64 the recommended route is the installer published on the project's releases page, named in the pattern CreatorHub-Setup with a version and a windows-x64 suffix. It installs WebView2 if it is missing, needs no Python or build step, and from version 0.2.0 checks for updates in the background at most once a day, only reminding rather than downloading.

Everyone else runs from source. Requirements are Python 3.10 or newer, a desktop environment because QR login needs a browser window, stable Google Chrome as an optional but recommended component, Node.js 18 or newer only if you enable the Xiaohongshu API publish compatibility mode, and system ffmpeg as an optional fallback that is otherwise supplied by a Python dependency.

Clone, then start:

bash
git clone https://github.com/3441293738/creatorhub.git
cd creatorhub

Windows uses the batch script, macOS and Linux the shell script:

bash
chmod +x start.sh
./start.sh

The first run creates a virtual environment, installs dependencies and Chromium, generates config.yaml, and opens the panel on 127.0.0.1 port 8000. The start script takes subcommands and flags, with the shell script substituted on Unix systems:

bash
.\start.cmd install        # 重新安装或更新依赖
.\start.cmd check          # 环境自检
.\start.cmd --no-open      # 启动后不自动打开页面
.\start.cmd --port 8080    # 使用其他端口
.\start.cmd --reload       # 开发模式

If you prefer to assemble it yourself, the manual path is a virtual environment, the requirements file, the browser component, a self-test, and the server:

bash
python -m venv .venv
python -m pip install -r requirements.txt
python -m patchright install chromium
python selftest.py
python -m uvicorn app.main:app --host 127.0.0.1 --port 8000

The Node.js dependency install exists only for the Xiaohongshu API publish compatibility mode.

Monitoring, collection and what the numbers mean

Keyword bulk collection is a one-off batch job, separate from continuous post monitoring. A single run accepts up to 20 keywords, with settings for posts per keyword, comments per post, whether second-level comments are included, and whether media is downloaded. Finished tasks can be reconfigured and re-run with deduplication against what was already collected, and results export to Excel.

Post monitoring takes a creator page, a work link, a short link or a platform identifier and records new works as they appear. Comment monitoring can subscribe to a single work or watch comments across an account's recent posts. Danmaku monitoring is separate from the comment section: it probes progressively along the video timeline, supports filters on time range, keywords, text length, like count and a capacity ceiling, and persists to SQLite, with your own videos going through the creator centre and public videos through playback interception.

One caveat is stated directly and deserves repeating: search results depend on platform ranking, login state and current visibility, so the post and comment counts you configure are collection ceilings, not the platform's full data. Downloads support resumable transfer and retry, Douyin offers quality selection, and Xiaohongshu handles both image sets and video.

Publishing, forwarding and the automated replies

Publishing differs by platform. Xiaohongshu supports image sets, video and scheduled posting, while Douyin, Kuaishou and WeChat Channels publish through their respective creator platforms. Cross-platform forwarding is a headline feature: a downloaded Douyin work can be forwarded to Xiaohongshu or Channels, and a Xiaohongshu work can be forwarded to Douyin, with title, body and topics editable before release.

Direct message auto-reply has more guardrails than most tools in this category. Rules support keywords, exclusion words, multiple templates, a reply cooldown and a maximum message age. Replies are generated as drafts pending review by default, and even with review disabled they remain subject to active hours, a unified write interval, hourly and daily caps, and a risk cooldown. The README also states that replies go through the account's own visible Chrome session and the page input box rather than assembled signed requests, and that uncertain submissions are not retried automatically.

For Douyin, own-account functions can be switched individually between api, browser and hybrid channels. The api mode opens no browser, hybrid falls back only when a direct call returns nothing usable, and a failure does not clear the existing snapshot.

Three things to weigh before you install

Licensing comes first. There is no LICENSE file in the repository tree and GitHub reports no licence for it, so there is no grant of rights to use, modify or redistribute the code. For a personal tool that may not matter, and for anyone thinking about building on it or shipping it inside a company, it is a blocker rather than a footnote.

Platform terms come second. This tool automates logging in, monitoring, downloading and republishing on consumer platforms, and forwarding someone else's work to another platform raises copyright questions that no tool can answer for you. The guardrails described above, draft-first replies, rate caps, active hours, exist because unrestrained automation on these platforms gets accounts restricted.

Maintenance coupling comes third. Everything here drives browser flows on platforms that change their front ends without notice. The release history shows the shape of that work: version 0.1.0 included fixes for Douyin sync chains broken by a front-end move to asynchronous hydration, and version 0.2.1 reworked Douyin account, follower and message sync across API, browser and hybrid modes. Keeping this running is an ongoing commitment.

Upkeep, and the narrower alternatives

The project is active. The last push was on 2026-09-17, and the three published releases are 0.2.1 on 2026-09-16, 0.2.0 on 2026-09-14 and 0.1.0 on 2026-09-10, a cadence of roughly one release a week. Version 0.2.0 added the background update checker, a Chinese-language installer, incremental update packages with resumable downloads, and backup of configuration and the database before install. Version 0.2.1 focused on the Douyin dual API and browser channel and repackaged the desktop build.

The obvious narrower alternative is doing any single job with a dedicated tool: yt-dlp for downloading shared links, which the project already depends on, or a platform's own creator centre for scheduling and analytics. What you give up is the unified view and the cross-platform forwarding. What you gain is a smaller surface that a front-end change cannot break.

The other comparison is a hosted social media scheduler. Those run on someone else's infrastructure and usually support the international platforms rather than these four, which for this audience is the deciding difference, since CreatorHub's coverage is entirely domestic Chinese platforms and its data never leaves your machine.

Editorial conclusion

Adopt it if you run several accounts across these platforms and want one dashboard for watching posts and comments, pulling media and republishing it elsewhere, with everything stored on your own machine. Do not adopt it if you need a licence grant, since there is no LICENSE file in the tree and GitHub reports no licence, or if your workflow depends on Xiaohongshu keyword collection, which is still listed as planned. Check three things first: that you are comfortable with the platform terms you are operating under when you forward someone else's post, that your machine can run a visible Chrome per account, and that you are on Windows x64 or willing to run from source on Python 3.10 or newer.

Frequently asked questions

Which platforms does CreatorHub support, and what is missing?

Douyin, Xiaohongshu, Kuaishou and WeChat Channels. WeChat Channels is the most limited, covering QR login, publishing and your own works, data and comments, but not monitoring or downloading other people's content. Keyword bulk collection is Douyin only, with the Xiaohongshu version listed as planned.

Does CreatorHub store my data on its own servers?

No. The README states that account login state, the database and media files are all stored locally. The public preview on GitHub Pages uses desensitised sample data and cannot log in, scrape, download or publish.

What do I need installed to run CreatorHub from source?

Python 3.10 or newer, a desktop environment for QR login, and optionally stable Google Chrome, which is recommended for Xiaohongshu login. Node.js 18 or newer is only needed for the Xiaohongshu API publish compatibility mode, and ffmpeg is optional because a Python dependency supplies it.

Is CreatorHub safe to leave automating comments and replies?

It ships with limits rather than open automation. Auto-replies become drafts awaiting review by default, and even with review off they are subject to active hours, a unified write interval, hourly and daily caps, and a risk cooldown. Replies are sent through the account's own visible Chrome session, and uncertain submissions are not retried.

Official sources

  1. 3441293738/creatorhub on GitHub
  2. Issues
  3. Project website
  4. README
  5. Releases
Community notes

Community notes