pURLfy

English | 简体中文

The ultimate URL purifier.

Note

Do you know that the name "pURLfy" is a combination of "purify" and "URL"? It can be pronounced as pjuɑrelfaɪ.

🪄 Functionalities

Purify URL: Remove redundant tracking parameters, skip redirecting pages, and extract the link that really matters.

• ⚡ Fast: Purify URLs quickly and efficiently.
• 🪶 Lightweight: Zero-dependency; Minified script less than 4kb.
• 📃 Rule-based: Perform purification based on rules, making it more flexible.
• 🔄️ Async: Calling purify won't block your thread.
• 🔁 Iterative purification: If the URL still contains tracking parameters after a single purification (e.g. URLs returned by redirect rules), it will continue to be purified.
• 📊 Statistics: You can track statistics of the purification process, including the number of links purified, the number of parameters removed, the number of URLs decoded, the number of URLs redirected, and the number of characters deleted, etc.

🤔 Usage

🚀 Quick Start

Visit our demo page, or try it out with our Telegram bot @purlfy_bot!

// Somewhat import `Purlfy` class from https://cdn.jsdelivr.net/gh/PRO-2684/pURLfy@latest/purlfy.min.js
const purifier = new Purlfy({ // Instantiate a Purlfy object
    fetchEnabled: true,
    lambdaEnabled: true,
});
const rules = await (await fetch("https://cdn.jsdelivr.net/gh/PRO-2684/pURLfy-rules@core-0.3.x/<ruleset>.json")).json(); // Rules
// You may also use GitHub raw link for really latest rules: https://raw.githubusercontent.com/PRO-2684/pURLfy-rules/core-0.3.x/<ruleset>.json
purifier.importRules(rules); // Import rules
const additionalRules = {}; // You can also add your own rules
purifier.importRules(additionalRules);
purifier.addEventListener("statisticschange", e => { // Add an event listener for statistics change
    console.log("Statistics increment:", e.detail); // Only available in platforms that support `CustomEvent`
    console.log("Current statistics:", purifier.getStatistics());
});
purifier.purify("https://example.com/?utm_source=123").then(console.log); // Purify a URL

Here's a list of test URLs that you can use to test pURLfy:

• Bilibili's short link: https://b23.tv/SI6OEcv
• Ordinary Tieba link: https://tieba.baidu.com/p/7989575070?share=none&fr=none&see_lz=none&share_from=none&sfc=none&client_type=none&client_version=none&st=none&is_video=none&unique=none
• MC Wiki's external link: https://link.mcmod.cn/target/aHR0cHM6Ly9naXRodWIuY29tL3dheTJtdWNobm9pc2UvQmV0dGVyQWR2YW5jZW1lbnRz
• Bing's search result: https://www.bing.com/ck/a?!&&p=de70ef254652193fJmltdHM9MTcxMjYyMDgwMCZpZ3VpZD0wMzhlNjdlMy1mN2I2LTZmMDktMGE3YS03M2JlZjZhMzZlOGMmaW5zaWQ9NTA2Nw&ptn=3&ver=2&hsh=3&fclid=038e67e3-f7b6-6f09-0a7a-73bef6a36e8c&psq=anti&u=a1aHR0cHM6Ly9nby5taWNyb3NvZnQuY29tL2Z3bGluay8_bGlua2lkPTg2ODkyMg&ntb=1
• A URL nested too many times that cannot be opened normally: https://www.minecraftforum.net/linkout?remoteUrl=https%3A%2F%2Fwww.urlshare.cn%2Fumirror_url_check%3Furl%3Dhttps%253A%252F%252Fc.pc.qq.com%252Fmiddlem.html%253Fpfurl%253Dhttps%25253A%25252F%25252Fgithub.com%25252Fjiashuaizhang%25252Frpc-encrypt%25253Futm_source%25253Dtest

📚 API

Constructor

new Purlfy({
    fetchEnabled: Boolean, // Enable the redirect mode (default: false)
    lambdaEnabled: Boolean, // Enable the lambda mode (default: false)
    maxIterations: Number, // Maximum number of iterations (default: 5)
    statistics: { // Initial statistics
        url: Number, // Number of links purified
        param: Number, // Number of parameters removed
        decoded: Number, // Number of URLs decoded (`param` mode)
        redirected: Number, // Number of URLs redirected (`redirect` mode)
        visited: Number, // Number of URLs visited (`visit` mode)
        char: Number, // Number of characters deleted
    },
    log: Function, // Log function (default is using `console.log` for output)
    fetch: async Function, // Function to fetch the given URL, should at least support `method`, `headers` and `redirect` in `options` parameter (default is using `fetch`)
})

Instance Methods

• importRules(rules: object): void: Import rules.
• purify(url: string): Promise<object>: Purify a URL.
  • url: The URL to be purified.
  • Returns a Promise that resolves to an object containing:
    • url: string: The purified URL.
    • rule: string: The matched rule.
• clearStatistics(): void: Clear statistics.
• clearRules(): void: Clear all imported rules.
• getStatistics(): object: Get statistics.
• addEventListener("statisticschange", callback: function): void: Add an event listener for statistics change.
  • The callback function will receive an CustomEvent / Event object based on whether the platform supports it.
  • If platform supports CustomEvent, the detail property of the event object will contain the incremental statistics.
• removeEventListener("statisticschange", callback: function): void: Remove an event listener for statistics change.

Instance Properties

You can change these properties after instantiation, and they will take effect for the next call to purify.

• fetchEnabled: Boolean: Whether the redirect mode is enabled.
• lambdaEnabled: Boolean: Whether the lambda mode is enabled.
• maxIterations: Number: Maximum number of iterations.

Static Properties

• Purlfy.version: string: The version of pURLfy.

📖 Rules

Community-contributed rules files are hosted on GitHub, and you can find them at pURLfy-rules. The format of the rules files is as follows:

{
    "<domain>": {
        "<path>": {
            // A single rule
            "description": "<规则描述>",
            "mode": "<模式>",
            // Other parameters
            "author": "<作者>"
        },
        // ...
    },
    // ...
}

✅ Path Matching

<domain>, <path>: The domain and a part of path, such as example.com/, /^.+\.example\.com$, path/ and page. Here's an explanation of them:

• The basic behavior is like paths on Unix file systems.
  • If not ending with /, its value will be treated as a rule.
  • If ending with /, there's more paths under it, like "folders" (theoretically, you can nest infinitely)
  • / is not allowed in the middle of <domain> or <path>.
• Note that if it starts with /, it will be treated as a RegExp pattern.
  • For example, /^.+\.example\.com$ will match all subdomains of example.com, and /^\d+$ will match a part of path that contains only digits.
  • Do remember to escape \, . etc in JSON strings.
  • Using RegExp is not recommended unless necessary, since it will slow down the matching process.
• If it's an empty string "", it will be treated as a FallBack rule: this rule will be used when no other rules are matched at this level.
• If there's multiple rules matched, the best matched rule will be used. (Exact match > RegExp match > FallBack rule)
• If you want a rule to match all paths under a domain, you can omit <path>, but remember to remove the / after the domain.

A simple example with comments showing the URLs that can be matched:

{
    "example.com/": {
        "a": {
            // The rule here will match "example.com/a"
        },
        "path/": {
            "to/": {
                "page": {
                    // The rule here will match "example.com/path/to/page"
                },
                "/^\\d+$": { // Remember to escape `\`
                    // The rule here will match all paths under "example.com/path/to/" that are composed of digits
                },
                "": {
                    // The rule here will match "example.com/path/to", excluding "page" and digits under it
                }
            },
            "": {
                // The rule here will match "example.com/path", excluding "to" under it
            }
        },
        "": {
            // The rule here will match "example.com", excluding "path" under it
        }
    },
    "example.org": {
        // The rule here will match every path under "example.org"
    },
    "": {
        // Fallback: this rule will be used for all paths that are not matched
    }
}

Here's an erroneous example:

{
    "example.com/": {
        "path/": { // Path ending with `/` will be treated as a "directory", thus you should remove the trailing `/`
            // Attempting to match "example.com/path"
        }
    },
    "example.org": { // Path not ending with `/` will be treated as a rule, thus you should add a trailing `/`
        "page": {
            // Attempting to match "example.org/page"
        }
    },
    "example.net/": {
        "path/to/page": { // Can't contain `/` in the middle - you should nest them
            // Attempting to match "example.net/path/to/page"
        },
        "/^\d+$": { // `\d` won't parse correctly in JSON strings, so use `\\d` instead
            // Attempting to match all paths under "example.net/" that are composed of digits
        }
    }
}

📃 A Single Rule

Paths not ending with / will be treated as a single rule, and there's multiple modes for a rule. The common parameters are as follows:

{
    "description": "<Rule Description>",
    "mode": "<Mode>",
    // Mode-specific parameters
    "author": "<Author>"
}

This table shows supported parameters for each mode:

Param\Mode	white	black	param	regex	redirect	visit	lambda
std	✅	✅	❌	❌	❌	❌	❌
params	✅	✅	✅	❌	❌	❌	❌
acts	❌	❌	✅	❌	❌	✅	❌
regex	❌	❌	❌	✅	❌	❌	❌
replace	❌	❌	❌	✅	❌	❌	❌
ua	❌	❌	❌	❌	✅	✅	❌
lambda	❌	❌	❌	❌	❌	❌	✅
continue	❌	❌	✅	✅	✅	✅	✅

🟢 Whitelist Mode white

Param	Type	Default
params	string[]	Required
std	Boolean	false

Under Whitelist mode, only the parameters specified in params will be kept, and others will be removed. std is for controlling whether the URL search string shall be deemed standard. Only if it is true or the URL search string is indeed standard will the URL be processed. Usually this is the most commonly used mode.

🔴 Blacklist Mode black

Param	Type	Default
params	string[]	Required
std	Boolean	false

Under Blacklist mode, the parameters specified in params will be removed, and others will be kept. std is for controlling whether the URL search string shall be deemed standard. Only if it is true or the URL search string is indeed standard will the URL be processed.

🟤 Specific Parameter Mode param

Param	Type	Default
params	string[]	Required
acts	string[]	["url"]
continue	Boolean	true

Under Specific Parameter mode, pURLfy will:

1. Attempt to extract the parameters specified in params in order, until the first existing parameter is matched.
2. Decode the parameter value using the processors specified in the acts array in order (if any acts value is invalid or throws an error, it is considered a failure and the original URL is returned).
3. Use the final result as the new URL.
4. If continue is not set to false, purify the new URL again.

🟣 Regex Mode regex

Param	Type	Default
regex	string[]	Required
replace	string[]	Required
continue	Boolean	true

Under Regex mode, pURLfy will, for each regex-replace pair:

1. Match the RegExp pattern specified in regex against the URL.
2. Replace all matched parts with the "replacement string" specified in replace.

If you'd like to learn more about the syntax of the "replacement string", please refer to the MDN documentation.

🟡 Redirect Mode redirect

Caution

For compatibility reasons, the redirect mode is disabled by default. Refer to the API documentation for enabling it.

Param	Type	Default
ua	string	undefined
continue	Boolean	true

Under Redirect mode, pURLfy will call constructor parameter fetch to get the redirected URL, by firing a HEAD request using ua to the matched URL and return the Location header or the updated response.url. If continue is not set to false, the new URL will be purified again.

🟠 Visit Mode visit

Caution

For compatibility reasons, the redirect mode is disabled by default. Refer to the API documentation for enabling it.

Param	Type	Default
ua	string	undefined
acts	string[]	["regex:<url_pattern>"]
continue	Boolean	true

Under Visit mode, pURLfy will visit the URL with ua, and if the URL has not beed redirected, it will call the processors specified in acts in order (<url_pattern> is https?:\/\/.(?:www\.)?[-a-zA-Z0-9@%._\+~#=]{2,256}\.[a-z]{2,6}\b(?:[-a-zA-Z0-9@:%_\+.~#?!&\/\/=]*)). The initial input to acts is of type string, i.e. the text returned by visiting the URL. If the URL has been redirected, the redirected URL will be returned. If continue is not set to false, the new URL will be purified again.

🔵 Lambda Mode lambda

Caution

For security reasons, the lambda mode is disabled by default. If you trust the rules provider, refer to the API documentation for enabling it.

Param	Type	Default
lambda	string	Required
continue	Boolean	true

Under Lambda mode, pURLfy will try to execute the lambda function specified in lambda and use the result as the new URL. The function shall be async, and its body should accept a single URL parameter url and return a new URL object. For example:

{
    "example.com": {
        "description": "示例",
        "mode": "lambda",
        "lambda": "url.searchParams.delete('key'); return url;",
        "continue": false,
        "author": "PRO-2684"
    },
    // ...
}

If URL https://example.com/?key=123 matches this rule, the key parameter will be deleted. After this operation, since continue is set to false, the URL returned by the function will not be purified again. Of course, this is not a good example, because this can be achieved by using Blacklist mode.

🖇️ Processors

Some processors support parameters, simply append them to the function name separated by a colon (:): func:arg1:arg2...:argn. The following processors are currently supported:

• url: string->string, URL decoding (decodeURIComponent)
• base64: string->string, Base64 decoding (decodeURIComponent(escape(atob(s.replaceAll('_', '/').replaceAll('-', '+')))))
• slice:start:end: string->string, String slicing (s.slice(start, end)), start and end will be converted to integers
• regex:<regex>: string->string, regex matching, returns the first match of the regex or an empty string if no match is found
• dom: string->Document, parse the string as a HTML Document object
• sel:<selector>: Any->Element/null, select the first element using CSS selector <selector> (The input shall have querySelector method)
• attr:<attribute>: Element->string, get the value of the attribute <attribute> of the element (getAttribute)
• text: Element->string, get the text content of the element (textContent)

😎 Projects Using pURLfy

Tip

If you are using pURLfy in your project, feel free to submit a PR to add your project here!

• Our Demo Page
• Our Telegram Bot @purlfy_bot (Source code)
• pURLfy for Tampermonkey
• LiteLoaderQQNT-pURLfy

🎉 Acknowledgments

• Thanks to Tarnhelm for providing some rules and for the initial inspiration of pURLfy.
• Thanks to this script on GreasyFork for providing some rules for pURLfy.
• Thanks to SearXNG for providing Bing's purification rule.
• The logo of pURLfy is a combination of the "Incognito" icon and the "Ghost" icon from SVG Repo.