# qd-today/templates: a public library of HAR templates for a check-in framework

> This repository is the public template library for the QD framework, a set of HAR files that describe how to perform a site's sign-in or check-in, contributed through issues and reviewed by maintainers. It is presented as a learning resource for the framework rather than a service, and the project states that templates are for study and not for commercial use.

**qd-today/templates** — 基于开源新版 QD 框架站发布的公共har模板库，仅供示例

- Repository: https://github.com/qd-today/templates
- Stars: 1,631 · Forks: 337
- Language: Unknown
- License: not declared
- Published: 2026-10-16 · Updated: 2026-10-16 · Language: en
- Canonical page: https://hysenlabs.com/projects/qd-today-templates

## What a template in this repository is

The repository is described in one line as a public template library published alongside the open source QD framework site, for example purposes only, and the README explains the relationship in the same way. The project is based on the open source QD framework, and its purpose is to guide users in learning how to use that framework. The templates in the project are all open source templates, for study and reference only, and the README asks that they not be used for commercial purposes.

The format is HAR, the HTTP Archive format, which is a JSON record of an HTTP request and response. A template here is therefore a description of a request to a particular site, complete with the placeholders the framework fills in at run time, such as a cookie or a token that the user obtained by logging in.

That is the whole design, and it explains the shape of the repository. The root of the repository is a very long flat list of HAR files, one per site, named after the site in a mixture of English and Chinese, with names like a cloud drive share site, a forum sign-in, a music site and a hardware forum, alongside the odd one with a bracket in it.

The framework does the work at run time, and the templates are the configuration. So the library is a corpus of examples rather than a program, and the README is explicit that it teaches the format.

## The framework lineage, and which one the templates target

The framework introduction section is short and it deals with the thing that trips new users up, which is that there are two projects with the same idea and different names.

The original framework was binux/qiandao, and it has stopped being maintained. The current one is qd-today/qd, which is the latest version of the framework and is still being updated. The templates in this repository are built against the latest framework, and the README says to work out the usage yourself from that project.

There is a compatibility caveat, and it is specific. A very small number of templates are not compatible with the old framework, and the instruction is to use the latest one if you hit that.

The other compatibility note is about a request scheme rather than a template. Only QD framework versions from 20211228 onwards support api:// requests. For any other version, the instruction when using a template is to upgrade the QD framework yourself, or to change api:// to http://localhost/.

So the two compatibility facts are the same kind of fact. The library targets a moving target, and the framework version you run determines which parts of the template vocabulary work.

## Templates are published by issue, and the automation around it

The publishing process is unusual and it is the part of the repository that says most about how it is run.

The README says it is no longer recommended to publish or update templates by pushing or opening a pull request, and asks instead that you do it through an issue, so that templates can be reviewed and managed automatically.

The steps are enumerated. You go to the issue interface, click the New issue button in the top right and choose the Get Started option for the template publishing type. You put the template name and the comment section in the title, fill in the content in the required format and submit. Then GitHub Actions runs and opens a pull request automatically, the repository owner reviews it, and once it is merged the template is published.

The tips underneath explain the machinery. You should not post anything else in the issue, or it will be closed. To update a template automatically, you edit the content of the existing issue rather than opening a new one. If you are not the issue author, you post the template content in the comments and mention the issue author or a repository administrator. To delete an invalid template automatically, you mark the corresponding issue label as invalid. And when the administrator merges the pull request, the recommended method is a squash merge.

So the issue tracker is the database and the pull request is the delivery mechanism. The separation exists so that a template is reviewed before it lands, and so that a bad template can be withdrawn by relabelling rather than by a revert.

## The template format, and the third-party library schema

The section on how to register a third-party library is the most technical part of the README, and it describes a self-hosted mirror of the template list.

Registration of third-party libraries has been open since the 20211021 version, and the default library provided is this repository. If you want to host your own, there are conditions. The repository root must contain a file called tpls_history.json, and the README gives the required shape.

```JSON
{
    "version":"版本号 yyyymmdd",
    "har": {
        "必填，和name保持一致，注意要在文件里保持唯一": {
            "name": "必填",
            "author": "选填，作者",
            "url": "选填，har链接",
            "update": false,
            "comments": "选填，har文件的注释，可用来解释har所需变量的说明",
            "filename": "必填，content为空时通过此来读取har",
            "content": "选填,不填则根据 filename 的值来读取对应的har文件,默认为base64编码",
            "date": "必填， 日期",
            "version":"必填， 版本号 yyyymmdd，框架通过版本号来判断是否更新模板",
            "commenturl":"选填，模板对应的评论区，留空时不显示按钮"
        }
    }
}
```

The comments field in each entry is the annotation that tells the reader which variables the template needs, which is the part that turns a raw request into something a user can fill in.

The other two conditions are about delivery. The acceleration service is jsdelivr by default and can be changed to ghproxy or fastgit through the configuration or an environment variable, and that only supports GitHub repositories. And the update rule is that an update is only accepted more than 24 hours after the previous one, that the presence of a template is judged by its name, that a new one is added directly if the name is not present, and that an existing one is updated only if its version is greater than the cached version.

The project adds a warning that it is not responsible for the content of any third-party library and that you should judge whether it is trustworthy yourself.

## The caveats: user agents, cookies and responsibility

The project statement is short and it puts the responsibility on the user, which is worth reading before you use any of this.

Any problems that arise from using the public templates are yours to bear. The publisher posts templates for demonstration only, and they must not be used for commercial purposes in any form. The project and its content are for learning and exchange, and it accepts no responsibility. And if you believe content in a template infringes your rights, you are asked to contact the repository administrators or the template publisher by email or through the discussions, and mention them, so it can be removed.

The first FAQ entry is the one that will actually bite a user, and it is a good example of a template library running into a website's defences.

The question is why some templates, after subscribing, report that you are not logged in when you supply the site's cookie. The answer is that it is not the template's fault, but that the site may validate the user agent, so a cookie obtained with one user agent stops working when the user agent changes. The instruction is to look at the user agent the template uses, and to log in with that user agent to obtain the cookie, and the note mentions a user agent switching extension for Firefox and the equivalent elsewhere.

That is a real operational detail of the format. A HAR template captures a request as one client made it, and if the site binds the session to the client fingerprint, the template is only half the story and the user has to supply the other half.

For anyone learning the format, the README points to a video demonstration and to a template writing specification.

## The state of the repository

Two practical notes about the project as it stands.

The first is the shape of the content. The repository is essentially a large flat list of HAR files with the documentation in the README, and the list is the product. There is no code to build, no test suite to run and no release to cut, and the templates are updated in place as the framework or the target sites change. The last push was on 2026-10-17, two days before this was written, and there are no tagged releases, which is consistent with a repository whose content is the deliverable and whose changes are individual files.

The second is the scale of the contribution. The README carries a long acknowledgement list of individual contributors, in no particular order, with a mixture of GitHub handles and a couple of names in Chinese, and the tip at the end that the list is incomplete. For a project of this kind, that list is the important artefact, because the templates come from the people who use the sites and know what a working request looks like for a service that no documentation describes.

There is also a QQ group for template exchange and bug reporting, and the README is explicit that it is only for that purpose, which is a reasonable boundary for a project that is a collection of other people's requests to other people's servers.

## Conclusion

Adopt this template library if you are learning the QD framework, or if you run a personal check-in setup and want working examples of the template format, because the repository exists to demonstrate the format and to be contributed to through issues. Do not treat it as a commercial service or a source of templates for a product, since the publisher restricts use to learning and demonstration. Before you rely on a template, read the entry for the site it targets, and expect to need the matching user agent, because several templates are rejected by sites that validate it.

## FAQ

### What is the qd-today/templates repository?

It is the public template library published alongside the open source QD framework, containing HAR template files that describe how to perform a site's sign-in or check-in. The project exists to guide users in learning how to use the framework, and the templates are open source, for study and reference, and must not be used for commercial purposes.

### How do I publish a template to the library?

Through an issue rather than a pull request. Open a new issue, choose the template publishing option, fill in the content in the required format and submit. GitHub Actions then opens a pull request automatically, and a repository owner reviews it and merges it, with a squash merge recommended.

### Which version of the QD framework do the templates work with?

The templates are built against the current framework, qd-today/qd, since the original binux/qiandao project has stopped being maintained. Only framework versions from 20211228 onwards support api:// requests, and for older versions you are asked to upgrade the framework or to change api:// to http://localhost/.

### Why does a template report that I am not logged in when I supply a cookie?

Because the site may validate the user agent, so a cookie obtained with one user agent stops working when the user agent changes. The advice is to check which user agent the template uses and to log in with that user agent to obtain the cookie, using a user agent switching extension if needed.

## Sources

- [Issues](https://github.com/qd-today/templates/issues)
- [qd-today/templates on GitHub](https://github.com/qd-today/templates)
- [README](https://github.com/qd-today/templates/blob/master/README.md)

---

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