Generate RSS feeds for web pages which don't offer their own.
You can also use the resulting feed with other services which take RSS as input, such as our Feed Control service.
Examples: Chomsky.info, advanced mode
Subscribe
Service shortcuts
Need more from the feed? Give it to one of our other tools:
Date: {{ item.date }}
{{ item.description }}
Don't want to rely on us? Comfortable running a PHP application on your own? You can host this yourself on your own server (one with at least PHP 7.2 installed).
When making HTTP requests, you can pass the following parameters to extract.php (either in a GET request or POST request). For most uses, supplying the page URL (url) and a selector (either in_id_or_class or item) is enough.
Example:
The form above uses the same parameters, so if you're unsure how to construct a URL, use the form and then examine the RSS feed URL that's produced.
| Parameter | Value | Description |
|---|---|---|
| url | string (URL) | Web page URL containing links which we want to extract for our RSS feed. This is required. |
| in_id_or_class | string (attribute value) | Find links inside elements whose id or class attribute matches this value. This translates to the following XPath: //a[@href and ancestor-or-self::*[@id="string" or contains(concat(" ",normalize-space(@class)," "), " string ")]] |
| item | string (CSS selector) | Look inside element(s) matching this CSS expression (for example: div.news .item). Cannot be used in combination with in_id_or_class. |
| item_title | string (CSS selector) or 0 | Extract item title from element matching CSS selector. This is applied within the context of elements selected by item. If omitted, the text of the first matching <a> element will be used. If set to 0, titles will not be included in the output. To use an element's attribute value rather than text content, use @attr, for example: 'img @alt'. To select the context element itself (element selected by item), pass ':scope'. |
| item_url | string (CSS selector), 0 or 1 | Extract item URL from element matching CSS selector. This is applied within the context of elements selected by item. If omitted, the URL of the first matching <a> element will be used. If set to 0, URLs will not be included in the output. If set to 1, all item URLs will point to the input URL. To use an attribute value other than 'href', use @attr, for example: 'img @src'. |
| item_desc | string (CSS selector) | Extract item description from element matching CSS selector. This is applied within the context of elements selected by item. If omitted, the generated feed will not include item descriptions. To use an element's attribute value rather than text content, use @attr, for example: 'img @alt'. To select the context element itself (element selected by item), pass ':scope'. |
| item_date | string (CSS selector) | Extract item date from element matching CSS selector. This is applied within the context of elements selected by item. If omitted, the generated feed will not include item date. To get the date from an attribute value rather than text content of the element, use @attr, for example: 'time @datetime'. |
| item_date_format | string | If the date is not recognised correctly (e.g. treated as US format instead of European), you can specify the format it appears in here. This should follow the createFromFormat pattern, e.g. 'j-M-Y'. |
| item_image | string (CSS selector) | Extract item image URL from element matching CSS selector. This is applied within the context of elements selected by item. If omitted, no image will be included in the output. Examples: 'img', 'img.main', 'img @data-src' (extracts URL from data-src attribute instead of src) |
| guid | 0 (default), 'url_title', 'url', 'title' | No guid field is included in the output by default. Pass 'url' to include a guid generated from the item URL, 'title' to have it generated from the item title, or 'url_title' use a combination of both. |
| url_contains[] | string (URL substring) | Filter out any item whose URL does not contain one of these substrings. This can appear multiple times. Example: /article/ |
| text_contains[] | string (word or phrase) | Filter out any item whose title or description does not contain any of the supplied words or phrases. This can appear multiple times. |
| since | number | Only include items published in the last x days. This is based on the date of the item as published by the page and selected with the date parameter. If this parameter is supplied and the items selected do not have a date, they will not be included in the output. |
| unique_url | 1 (default), 0 | If multiple matching items have the same URL, only the first encountered will be kept. Set to 0 to keep all matching URLs. Note: this is enabled by default, but always disabled when you ask for URLs to be excluded from output (item_url=0). |
| unique_title | 1 (default), 0 | If multiple matching items have the same title, only the first encountered will be kept. Set to 0 to keep all matching titles. Note: this is enabled by default, but always disabled when you ask for titles to be excluded from output (item_title=0) |
| strip | string (CSS selector) | Remove elements matching CSS selector. This will be processed before we start looking for items. To strip multiple elements, use commas. For example: .header,.footer |
| keep_qs_params | 1 (default), 0, param1,param2 | Determines the query string parameters that get preserved in item URLs. 1 = all preserved, 0 = none preserved. You can also pass individual parameters by name, e.g. 'articleId' to only preserve that parameter. Use a comma to supply more than one, e.g. 'id,category'. |
| strip_if_url[] | string | Filter out any item whose URL contains any of the supplied words or phrases. This can appear multiple times. |
| strip_if_text[] | string | Filter out any item whose title or description contain any of the supplied words or phrases. This can appear multiple times. |
| feed_title | string | The feed title to use in the generated feed. If omitted, we'll use whatever's in the <title> element of the web page requested. Note: this should be the actual title, not a selector. |
| max | number (limit: 20) | The maximum number of items to return. The limit can be changed in the config file. |
| format | rss (default), json | By default the output will be formatted as RSS 2.0. Set this to JSON if you prefer that. |
| order | document (default), reverse | Results are returned in document order by default (ie. in the order they appear in the source HTML). Set this to 'reverse' to reverse it. |
| ua | string | HTTP User-Agent header. How Feed Creator should present itself when fetching content. If a site only produces content for certain browsers, you can use this field to identify as that browser. This is sent in the HTTP request in a 'User-Agent' header. For example: "Mozilla/5.0 (x64; rv:77.0) Gecko/20100101 Firefox/77.0". (If self hosting, the default can be changed in the config file.) |
| referer | string, 0 or 1 | HTTP Referer header. Set to 0 to disable sending a referer header. Set to 1 to use the page URL passed in url. Or specify a custom Referer header, e.g. "https://www.google.com". (If self hosting, the default can be changed in the config file.) |
| cookie | string | HTTP Cookie header. This should be used if the site needs you to be logged in to view content, or to bypass GDPR and cookie walls. For example: "euconsent=1; name2=value2;" |
| parser | html5 (default), libxml | Libxml is a slightly faster parser, but might not handle all pages correctly. If speed is important for you, try changing the parser to libxml - in most cases the results should be the same. |
| proxy | string | Self-hosted use only. If you add proxy servers to the config file, you can pass the name of the proxy server you want to use for this request, e.g. 'rotating-proxy-us'. |
| saved[] | string | Saved request parameters. In the config file (when self hosting), you can associate a set of request parameters with a given name. You can then pass that name in the request instead of the request parameters. This is useful if you have one set of parameters you want to apply to a number of generated feeds, or if you want to update parameters later without changing the generated feed URL. |
Required parameters: url must be supplied.
In addition to the parameters above, there's a configuration file which you can edit.
Features include:
To change the configuration, save a copy of config.php and save it inside the custom folder and make any changes you like to it.
If everything works fine, feel free to modify this page by following the steps below:
Next time you load this page, it will automatically load custom/index.php instead.
We have documentation if you need help. You can also check out the forum or email us at help@fivefilters.org.