linktre-tools: a link directory kept in a single README file

The repository is one README, and that is the whole architecture

The most important fact about PennyJoly/linktre-tools is the repository listing, which contains a single file, README.md. There is no source directory, no build script, no package manifest, and no test suite. The project is the content of a navigation site rather than the site itself. The README describes itself as the site repository for a navigation site called 节点导航站, at linktre.cc, and says it collects independent developer technical tools, AI products, cross-border e-commerce, and office assistant tools for content creators, with collection ongoing. The primary language is reported as unknown, which is consistent with a file that is mostly prose and markdown tables rather than code. For a reader, this reframes the question. You are not evaluating a library or a service here. You are evaluating a dataset with a contribution workflow, and the question is whether the entries and the workflow are worth your time.

Entries are markdown table rows, and the shape is consistent

The content is organised into sections, each with a two column table of site and description. The first is a recommended section with a short curated list, the second is a daily updates section, and the third collects popular overseas AI products. The format is uniform enough to be machine readable and human scannable at the same time. Each row carries a linked site name and a one line description, and the links point either at the product's own domain or at a `siteDetails` path on the navigation site itself, which is where the site's own catalogue pages live. The daily section in the current README mixes Chinese and English descriptions, which tells you the submitters are international even though the project's framing is aimed at Chinese-language users going out to sell products overseas. Contributors add rows by opening pull requests or issues, and the README asks for both, with the promise of daily attention to incoming submissions.

The NoFollow policy shapes what the list is for

One line in the README explains the submission rules better than any feature list would. It states that the collected sites are all NoFollow, and that if you need follow weight transferred you should add the author's WeChat account with a note asking for weight transfer. That is a blunt statement of the site's business model, and it tells you what kind of directory this is. Entries do not pass ranking signals to the sites they point at, so the list is a discovery aid rather than a link-building scheme, and the author is explicit that weight passing is a separate arrangement handled by direct contact. For a user browsing the catalogue, that is simply how it works. For anyone planning to reuse the list as a source of outbound links for their own site, it is a boundary rather than a footnote, and it applies to every entry without exception.

What the listings actually contain, and what they do not

Reading the entries, the catalogue covers a specific band of software. The recommended section includes a subscription and API gateway service for AI coding workflows, a Nuxt 3 based SaaS starter with payment, Google login, multi-language routing and SEO tooling, a curated directory of open source projects, a free temporary mailbox service, and the open source version of the navigation site itself along with the open source version of the Nuxt starter. The daily section covers AI document parsing and translation from PDF, scans and images into Word, Markdown and LaTeX with formula and table recognition, a multi-language catalogue of AI tools and models, an AI image prompt generator and prompt library, a PDF translation tool, an AI music generation platform, and a manga translation tool. What none of the entries include is a price, a licence, a changelog, or a verification statement. Each is a name, a URL, and a promotional sentence.

The popular AI products section is a familiar, dated reference list

The third table is a different kind of artefact. It lists established names such as ChatGPT, Midjourney, Claude, Le Chat, Gemma, Sora, Suno AI, Llama, SearchGPT, v0.dev, Udio, Notion AI, Grok, and a set of smaller tools including a free GPT detector built on Transformers, a TikTok to MP3 converter, an AI cartoon generator, a logo maker, a website analysis and A B testing tool, a 3D model to 64x64 pixel converter, a local AI cluster builder for home devices, and a code assistant. Many of these entries point at the navigation site's own `siteDetails` pages rather than to the vendors directly, which suggests the site maintains a record page for each listing. Two of the descriptions are worth noting for what they reveal about the source. One says only that the product is an AI tool, and another is a full paragraph about a free online AI photo editor. The variation in detail tells you the descriptions are written by submitters rather than by a reviewer.

The gaps: no licence, no code, and no way to check the entries

Three limits define what this repository can and cannot be used for. First, licensing. GitHub reports the licence as unknown, and the README does not state one, so the default copyright position applies, which means you cannot copy the list into your own project and redistribute it with confidence. Second, there is no code. The navigation site that consumes this data is a separate open source project, linked from the recommended section, and the open source versions of the site and the Nuxt starter are separate repositories with their own terms. The tooling repository here is content only. Third, and most important for a directory, there is no verification layer. The README says the collection continues and that submissions are welcome, and it asks readers to star the repository and to submit pull requests and issues, but nothing in the file indicates that a listed site was tested, that its description was checked, or that the entry is still current. A link directory is only as good as its curation, and curation is the part this repository does not show.

Who this is for, and what to check before you lean on it

The audience is fairly clear. A developer or a small content operation looking for a large seed list of AI tools and developer utilities, in Chinese, organised and updated frequently, gets value from this file. A person running their own directory site can use it as a starting catalogue and as a model for how the submission workflow is shaped. A company building a vetted marketplace of AI tools, or anyone with a procurement or legal requirement to know what a tool does and under what terms, gets nothing useful from a list of one-line marketing summaries with no licensing metadata. If you fall in the second group, the honest starting point is that this file is a list of leads. The concrete next step is to check the licence question for yourself, since the repository declares none, and to spot check a sample of the entries against the vendors' own pages before you rely on any of the descriptions, because the entries are written by the people who submitted them and the README makes no claim that they were verified.

Editorial conclusion

linktre-tools fits an independent developer or a content site operator who wants a ready-made catalogue of AI tools and developer utilities to seed a directory of their own, and who is comfortable working with data rather than code. It is the wrong source for anything requiring verified, maintained, or licensed entries, because the repository contains only a README, declares no licence, and describes entries in one-line marketing summaries. Before you depend on it, read the license question yourself, since GitHub reports the licence as unknown, and treat every listing as a lead to verify rather than a fact, because the descriptions come from the submitters and the directory's own policy is that outbound links carry no follow weight.

Frequently asked questions

Official sources

1. Issues
2. PennyJoly/linktre-tools on GitHub
3. README