# insoxin API: a self-hosted collection of public web APIs

> The insoxin/API project is a Docker-packaged collection of small public interfaces covering weather, wallpapers, QR codes, hot searches, music, video and utility tools, served from one origin with REST-style paths and JSON responses. It suits a developer who wants a couple of these endpoints without writing them, and its terms are stricter than the free-call promise suggests.

**insoxin/API** — API For Docker 一个基于多种编程语言开源免费不限制提供生活常用,出行服务,开发工具,金融服务,通讯服务和公益大数据的平台.

- Repository: https://github.com/insoxin/API
- Website: https://api.isoyu.com/
- Stars: 1,339 · Forks: 235
- Language: JavaScript
- License: AGPL-3.0
- Published: 2026-10-19 · Updated: 2026-10-19 · Language: en
- Canonical page: https://hysenlabs.com/projects/insoxin-api

## What the project actually is

The insoxin/API repository describes itself as a platform of open APIs built on several programming languages, free and with no limit on calls, covering daily life services, travel, developer tools, financial services, communication services and public-interest big data. The homepage is api.isoyu.com and the primary language recorded for the repository is JavaScript, though the description is explicit that the collection spans multiple languages, which is consistent with the endpoints doing very different things. The framing in the README is that this is a personal collection of APIs, gathered and maintained over nearly ten years, added to as needed, and that the page lists the interfaces currently public. That is a useful piece of context, because it tells you the collection is opportunistic rather than designed as a coherent product, and the endpoint list reads that way. There is one release, v0.1 dated 2022-05-19, and the last push was on 2026-09-05, so the code is still being touched even though nothing has been tagged in years. The licence is AGPL-3.0, which is the one that matters most here.

## The endpoint list is a catalogue of Chinese web scraping

The interface list is the best way to understand the project. It includes public-interest data such as the core data of the Baohui Jia (a missing-children organisation) and national wanted notices with SSL, plus AI analysis of mammogram photographs, Shanghai municipal data, and news feeds classified into headlines, military, entertainment, sports, technology, art, education and breaking news. There are video categories for quality videos, funny videos, beauty videos, sports video and news footage, image interfaces for cosplay and for a Huaban album, joke and weather interfaces, and developer-oriented feeds such as front-end development daily reports and Zhihu daily. Commercially it is thinner, with recruitment listings, ISBN lookup, IELTS score lookup and a CSGO account information query. The common thread is that each of these is a small fetch from a Chinese website, usually re-serialised as JSON, and the value of the project is that the parsing work is already done. The topics list confirms it, with imgur-api, img, aru, qq, proxy, url, voice, pay and pay-related entries.

## The interface design, in the project's own words

The README describes the design in a paragraph that is unusually specific. The service runs for a long time in Linux Docker containers, with data stored primarily in OceanBase. The interfaces use a REST style with predictable, resource-oriented request addresses, express success or failure through standard HTTP status codes, and use HTTP Authentication alongside standard request methods for usability. Every API request returns specification-friendly JSON including error information. The OceanBase detail is the interesting one for anyone thinking about running this themselves, because it is a distributed relational database rather than the MySQL or PostgreSQL a container would normally bring along, and it is a substantial piece of infrastructure to stand up for a collection of small fetchers. The HTTP Authentication piece is also unusual for a free public service, since it means the operator controls access rather than leaving every endpoint open.

## Running it in Docker

The installation is two steps, and the Dockerfile in the repository is a single line, `FROM insoxin/api`, which means the image is not built from this source tree but pulled from the published one. You pull the image:

```bash
docker pull insoxin/api:latest
```

Then you run the container:

```bash
docker run -dit --name api -p 8080:80  --restart unless-stopped insoxin/api:latest
```

After that you access it at `IP:8080`. The port mapping is the whole configuration surface, 8080 on the host to 80 in the container, and the restart policy means the service comes back after a reboot. The repository root also carries a file named in Chinese that marks itself as important and is an installation note, along with `runtime/`, `extend/`, `js/`, `gy/`, `css/`, `demo/` and `images/` directories, which suggests the container serves the documentation site and the API from the same image.

## The terms, which are stricter than the headline

Three notices in the README matter more than the feature list. The first is the data collection notice. The API records your domain, usage traffic and call counts and stores them for six months, and it says to not use it if you mind that, and it asks daily IPs over a thousand to apply to be whitelisted. The second is the ownership notice. The API holds all the operating rights, and in certain specific circumstances the operator has the right to prohibit your website from calling it. The third is the usage restriction. Illegal use is forbidden, commercial use requires authorisation, and if a large volume of prohibited abuse is found, the affected function gets an authorisation code applied, available only to users with Ministry of Industry and Information Technology filing or Public Security Bureau filing. That last clause is unusual and worth reading carefully, because it means the access rules for a given endpoint can change based on the operator's assessment of traffic from any given user. The repository root has a `COMMERCIAL.md` file, and the administrator contact is given as admin@isoyu.com.

## The documentation site, and the pieces of the repository

The documentation is a VitePress site in the `docs/` directory, and the package manifest at the root is the docs toolchain rather than the API, with the scripts `docs:dev`, `docs:build` and `docs:serve`, plus `dev`, `build` and `serve` aliases, and a devDependencies entry of vitepress and vue at the `latest` tag. You install the dependencies with `npm install`, then run `npm run docs:dev` for a local preview or `npm run docs:build` to generate the static site, whose output lands in `docs/.vitepress/dist`. The pnpm peer dependency rules ignore a missing `@algolia/client-search`. The rest of the tree is the public site that the container serves, with a `demo/index.html`, a `404.html`, a `contact.html`, a `help.html`, a `robots.txt`, a `sitemap.xml`, and the assets in `images/`, `fonts/`, `css/` and `js/`. The `extend/` and `runtime/` directories are where the additional endpoints and the server-side code sit, and there is a `gy.js` file and a `gy/` directory that look like the commercial or member-gated parts.

## What this is not

It is worth being blunt about the category, because the marketing language invites the wrong comparison. This is not an API platform in the sense of a gateway with documentation, keys, quotas and a support contract. It is a personal collection of scrapers for Chinese web services, exposed over HTTP, with a documentation site and a container image. There is no versioning story for individual endpoints, no deprecation policy, no per-endpoint uptime information, and the single v0.1 release from 2022 tells you the packaging is not where the maintenance effort goes. The AGPL-3.0 licence is the other point to check, since it is a copyleft licence with network use provisions, which matters if you plan to modify the endpoints and run them as a service of your own. The honest summary is that this is a useful reference implementation and a working instance of a set of small fetches, and it should be treated as infrastructure to inspect rather than a dependency to build a product on without a plan for what happens when an upstream site changes its markup.

## Conclusion

Adopt insoxin/API if you want a handful of small public interfaces behind one origin and one container, and you are willing to accept the terms it states, which include logging your domain, traffic and call counts for six months, and reserve the right to block a site from calling. Skip it if you need a supported service with an SLA, because the project explicitly disclaims any obligation and requires authorisation for commercial use. Before you build on it, read the terms and the COMMERCIAL.md file in the repository root, since the free and unlimited claim is qualified by the data collection notice, and note that the single release is v0.1 from 2022-05-19 while the last push was on 2026-09-05.

## FAQ

### How do I run the insoxin API in Docker?

Pull the image with `docker pull insoxin/api:latest`, then run `docker run -dit --name api -p 8080:80 --restart unless-stopped insoxin/api:latest` and access it at IP:8080. The Dockerfile in the repository is a single FROM insoxin/api line, so the image comes from the published registry rather than being built from source.

### Does the insoxin API log my usage?

Yes. The README states the API records your domain, usage traffic and call counts and stores them for six months, asks that you not use it if you mind that, and says daily IPs over a thousand should apply to be whitelisted.

### Can I use the insoxin API commercially?

Not without authorisation. The README states that commercial use requires authorisation, that illegal use is forbidden, and that if large-scale prohibited abuse is found the affected function gets an authorisation code that only users with Ministry of Industry and Information Technology filing or Public Security Bureau filing may apply for. There is a COMMERCIAL.md file in the repository root.

### What kinds of interfaces does the insoxin API provide?

Public-interest data, classified news and video feeds, image interfaces, weather, QR code generation and decoding, wallpapers, hot searches from Baidu, a Douyin watermark removal interface, an audio resource sniffer, a QQ nickname and avatar lookup, URL shortening, an ISBN lookup, an image NSFW check and assorted developer utilities. The list is described as a personal collection gathered over ten years.

### How is the insoxin API documentation built?

The docs are a VitePress site under the docs directory. Install dependencies with npm install, then run npm run docs:dev for a local preview or npm run docs:build to generate the static site into docs/.vitepress/dist.

## Sources

- [insoxin/API on GitHub](https://github.com/insoxin/API)
- [License: AGPL-3.0](https://github.com/insoxin/API/blob/master/LICENSE)
- [Project website](https://api.isoyu.com/)
- [README](https://github.com/insoxin/API/blob/master/README.md)
- [Releases](https://github.com/insoxin/API/releases)

---

Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/insoxin-api