OpenClaw 101: a Chinese-language resource index for OpenClaw, built as a Next.js site

What OpenClaw 101 actually is, and who it is written for

The repository describes itself as an open-source resource aggregation site for OpenClaw, a separate open-source AI personal assistant platform. That distinction matters more than it first appears. Cloning openclaw101 gives you a Next.js website whose job is to list and filter links. It does not install OpenClaw, it does not run an assistant, and it holds no model credentials. Anyone who arrives expecting a working assistant will be disappointed within the first minute.

The audience is narrow and clearly stated: Chinese-speaking users who want to get started with OpenClaw quickly. The README frames the site around a 7-day learning path that links out to a Feishu knowledge base, plus skill recommendations sorted by scenario. The collected sources lean toward Chinese cloud and community platforms, including Alibaba Cloud material on DingTalk and WeCom AppFlow, Tencent Cloud material on Feishu and WeCom integration, Bilibili video tutorials, and posts from Blog Garden and CSDN. English sources are present too, from DigitalOcean, Hostinger, Codecademy, IBM Think and Reddit, and the site claims bilingual coverage.

If you are an English-only reader, the value proposition shrinks. The interface and the framing are Chinese-first, the learning path lives in Feishu, and the README does not describe a full English translation of the guide itself. English entries in the resource list are links out, not translated content.

How the site is put together: one data file, eight components

The architecture is about as simple as a Next.js content site gets. The README's project structure shows src/app for routes (page.tsx for the home page, resources/page.tsx for the aggregation page, layout.tsx and globals.css), src/components for eight named components (Hero, WhatIs, LearningPath, Skills, ResourcesSection, Community, Navbar, Footer), and src/data/resources.ts as the single place where every resource lives.

That last file is the whole content model. Each resource is an object with title, desc, url, source, lang, category, featured and tags. Categories are a fixed set defined in the README: official, getting-started, channel-integration, skill-dev, video, deep-dive, tools and cloud-deploy. Filtering and search on the resources page operate over this array, so adding a resource means editing TypeScript, not touching a database or a CMS.

The dependency list confirms the shape. There is no database client, no auth library, no API layer. What you do see is MDX tooling (@mdx-js/loader, @mdx-js/react, @next/mdx), gray-matter for frontmatter, react-markdown with remark-gfm and rehype-raw, and reading-time. So the repository is set up for both a structured link index and longer markdown or MDX articles, even though the README only documents the link index workflow. The top-level content/ directory and the scripts/ and ops/ directories are not explained in the README, which is a real gap: a contributor reading only the README will not know what content/ holds or what the ops/ scripts do.

Running OpenClaw 101 locally and adding your first resource

The README gives a four-step quick start. Cloning, installing and starting the dev server are the standard Next.js flow, and npm run dev should serve the site on the default Next.js port, 3000. The README does not state the port explicitly, so treat that as the default rather than a documented setting.

git clone https://github.com/mengjian-github/openclaw101.git
cd openclaw101
npm install
npm run dev

Note the build script, because it is not a plain next build. The package.json defines build as next build followed by node scripts/postbuild-seo.mjs. That postbuild step runs after compilation and is not described in the README, so if you are wiring this into CI, check what that script writes before you assume the build output is only static HTML.

npm run build
npm start

To add a resource, edit src/data/resources.ts and append an object in the existing shape. The README's example uses these exact keys:

{
  title: '资源标题',
  desc: '一句话描述',
  url: 'https://...',
  source: '来源名称',
  lang: 'zh',
  category: 'getting-started',
  featured: false,
  tags: ['标签1', '标签2'],
}

The lang field takes 'zh' or 'en'. The category field must be one of the eight documented values, so a typo like 'getting_started' will not match the filter. After saving, the dev server hot-reloads and the new entry should appear on the resources page under its category. The contribution flow the README asks for is a fork, an edit to that file, and a pull request with a short description.

What it does not do, and where the documentation stops short

The most concrete limitation is that this is a link directory, and link directories rot. The README says resources are continuously being collected, but nothing in the repository layout suggests automated link checking. There is no test runner in package.json, no CI workflow described in the README, and no documented validation that a URL still resolves. A tutorial on Tencent Cloud about Feishu integration can be rewritten or removed without anything in this repository noticing.

The licensing situation is inconsistent. The README's final section states MIT License, but the repository metadata retrieved for this project lists the license as unknown, and there is no LICENSE file among the top-level entries (.github/, .gitignore, README.md, content/, next.config.js, ops/, package-lock.json, package.json, postcss.config.js, public/, scripts/, src/, tailwind.config.ts, tsconfig.json). If you plan to reuse the code, that discrepancy is worth resolving with the maintainer before you rely on the MIT claim.

There is also a stray top-level entry named = in the repository listing. It looks like an accidental artifact from a shell redirection. It is harmless, but it is a signal about how much cleanup the repository has had.

The README does not document deployment beyond naming Cloudflare Pages as the hosting platform, does not describe the postbuild SEO script, does not explain the content/ or ops/ directories, and does not state a Node version requirement. Anyone reproducing the build on an unfamiliar machine is working from the dependency versions alone.

OpenClaw 101 compared with a general-purpose awesome list

The obvious alternative is a plain awesome-style markdown list on GitHub. The difference in approach is structural. An awesome list is a single README rendered by GitHub, with no filtering, no per-item metadata and no way to mark a resource as recommended. OpenClaw 101 puts the same kind of curated links into typed objects and renders them through a Next.js application, which buys filtering by category, filtering by language, keyword search and a featured flag.

That trade is not free. A markdown list can be edited in the browser in thirty seconds by someone who has never run Node. Editing OpenClaw 101 requires cloning, installing dependencies and running a TypeScript build, or at minimum understanding the object shape well enough to submit a correct pull request. The contributor pool for a Next.js site is smaller than the contributor pool for a README, and this project's own README spends most of its contribution section explaining the object format, which suggests the maintainer knows that.

The second alternative, for the reader rather than the contributor, is the OpenClaw project's own documentation. If the upstream repository documents installation and channels, the index adds curation and Chinese-language discovery rather than new technical information. OpenClaw 101 is a map, and maps are only as good as the roads underneath them.

Maintenance, upgrade cost and the licence question

The last push to the default branch was on 2026-07-28. That is recent enough that the repository is not abandoned, but the README gives no release history and no retrieved releases exist, so there is no versioned changelog to read. Upgrades are therefore driven by the dependency tree rather than by project releases.

That tree carries the real maintenance cost. Next.js is pinned at 14.2.29 while @next/mdx is at ^16.1.6, a major-version gap between the framework and its MDX plugin. React is on 18.3.1. Tailwind is on 3.4.4. A future Next.js major upgrade will require touching next.config.js, the MDX wiring and possibly the app router files. For a site whose content is eight components and one data array, that is a manageable but recurring chore, and it is the kind of chore that quietly stops happening on small content sites.

On licensing, the README states MIT and describes it as free to use, modify and distribute. The repository metadata does not confirm a licence, and no LICENSE file appears in the top-level listing. MIT on a site like this would mean you can reuse the components and the data structure in your own project, but you would still need to respect the rights of the linked tutorial authors, whose content lives on Alibaba Cloud, Tencent Cloud, Bilibili and elsewhere and is not covered by this repository's licence at all. That is a factual boundary, not legal advice; if you intend to republish linked material, ask the original authors.

Editorial conclusion

Adopt OpenClaw 101 if you read Chinese and want a curated index of OpenClaw deployment and integration tutorials rather than a tool that runs on your machine. Skip it if you need English documentation, offline copies, or anything that executes an assistant. Verify first that openclaw101.dev resolves and that the sources you care about are still reachable, then check src/data/resources.ts for entries matching your cloud provider and messaging platform. If you only want to read, the deployed site is enough; fork only when you intend to add entries.

Frequently asked questions

Official sources

1. Issues
2. mengjian-github/openclaw101 on GitHub
3. Project website
4. README