1. Introduction

Newsboat is an RSS/Atom feedreader. RSS and Atom are a number of widely-used XML formats to transmit, publish and syndicate articles, for example news or blog articles. Newsboat is designed to be used on text terminals on Unix or Unix-like systems such as GNU/Linux, FreeBSD or macOS.

1.1. Platforms

Newsboat has been tested on Linux (with glibc and musl-libc), FreeBSD and macOS.

NetBSD is currently not supported, due to technical limitations in the iconv() implementation.

1.2. Why "Newsboat"?

"Newsboat" is a play on the name of its ancestor, "Newsbeuter". They’re spelled quite differently, but sound similar. ("Newsbeuter" is a pun on German word "wildbeuter"; "newsboat" is an English word.)

Newsboats were the vessels that collected and delivered news shuffling between boats in the port. Newsboat the program will collect the news for you, just like its namesakes did back in the day.

2. Installation

2.1. Pre-built binaries

Newsboat binaries are available in a number of repositories:

2.2. From source

2.2.1. Download the source code

The most up-to-date source code can always be downloaded from the Git repository:

$ git clone git://github.com/newsboat/newsboat.git

There are also signed release tarballs available from our website.

2.2.2. Install dependencies

Newsboat depends on a number of libraries, which need to be installed before Newsboat can be compiled. Make sure to install the header files as well (on Debian and derivatives, headers are in -dev packages, e.g. libsqlite3-dev.)

Developers will also need:

2.2.3. Compile and install

There are a few different ways:

  • build inside Docker. For that, check out doc/docker.md in the source tree. Note that the resulting binary might not run outside of that same Docker container if your system doesn’t have all the necessary libraries, or if their versions are too old;

  • build in a chroot: to avoid polluting your system with developer packages, or to avoid upgrading, you might use a tool like debootstrap to create an isolated environment. Once that’s done, just build from source as outlined in the next item;

  • build from source:

    $ make                   #  pass -jN to use N CPU cores, e.g. -j8
    $ sudo make install      #  install everything under /usr/local

    To install to a different directory, pass prefix like so: sudo make prefix=/opt/newsboat install.

    To uninstall, run sudo make uninstall.

Cross-compilers need to set CARGO_BUILD_TARGET; see cargo documentation.

3. First Steps

After you’ve installed newsboat, you can run it for the first time by typing newsboat on your command prompt. This will bring you the following message:

Error: no URLs configured. Please fill the file /home/ak/.newsboat/urls with RSS feed URLs or import an OPML file.

newsboat 2.10
usage: ./newsboat [-i <file>|-e] [-u <urlfile>] [-c <cachefile>] [-x <command> ...] [-h]
    -e, --export-to-opml            export OPML feed to stdout
    -r, --refresh-on-start          refresh feeds on start
    -i, --import-from-opml=<file>   import OPML file
    -u, --url-file=<urlfile>        read RSS feed URLs from <urlfile>
    -c, --cache-file=<cachefile>    use <cachefile> as cache file
    -C, --config-file=<configfile>  read configuration from <configfile>
    -X, --vacuum                    compact the cache
    -x, --execute=<command>...      execute list of commands
    -q, --quiet                     quiet startup
    -v, --version                   get version information
    -l, --log-level=<loglevel>      write a log with a certain loglevel (valid values: 1 to 6)
    -d, --log-file=<logfile>        use <logfile> as output log file
    -E, --export-to-file=<file>     export list of read articles to <file>
    -I, --import-from-file=<file>   import list of read articles from <file>
    -h, --help                      this help

This means that newsboat can’t start without any configured feeds. To add feeds to newsboat, you can either add URLs to the configuration file $HOME/.newsboat/urls or you can import an OPML file by running newsboat -i blogroll.opml. To manually add URLs, open the file with your favorite text editor and add the URLs, one per line:

http://rss.cnn.com/rss/cnn_topstories.rss
http://newsrss.bbc.co.uk/rss/newsonline_world_edition/front_page/rss.xml

If you need to add URLs that have restricted access via username/password, simply provide the username/password in the following way:

https://username:password@hostname.domain.tld/feed.rss

In order to protect username and password, make sure that $HOME/.newsboat/urls is only readable by you and, optionally, your group:

$ chmod u=rw,g=r,o= ~/.newsboat/urls

Newsboat also makes sure that usernames and passwords within URLs aren’t displayed in its user interface. In case there is a @ in the username, you need to write it as %40 instead so that it can be distinguished from the @ that separates the username/password part from the hostname part.

You can also configure local files as feeds, by prefixing the local path with file:// and adding it to the urls file:

file:///var/log/rss_eventlog.xml

Now you can run newsboat again, and it will present you with a controllable list of the URLs that you configured previously. You can now start downloading the feeds, either by pressing "R" to download all feeds, or by pressing "r" to download the currently selected feed. You can then select a feed you want to read, and by pressing "Enter", you can go to the article list for this feed. This works even while the downloading is still in progress.

You can now see the list of available articles by their title. A "N" on the left indicates that an article wasn’t read yet. Pressing "Enter" brings you to the content of the article. You can scroll through this text, and also run a browser (default: lynx) to view the complete article if the content is empty or just an abstract or a short description. Each URL in the article has a number next to it; to open it, type # and then the number, then press "Enter". For single-digit links, like 3, you can just press that number on the keyboard.

Pressing "q" brings you back to the article list, and pressing "q" again brings you back to the feed list. Pressing "q" a third time then closes newsboat.

Newsboat caches the article that it downloads. This means that when you start newsboat again and reload a feed, the old articles can still be read even if they aren’t in the current RSS feeds anymore. Optionally you can configure how many articles shall be preserved by feed so that the article backlog doesn’t grow endlessly (see max-items below).

Newsboat also uses a number of measures to preserve the users' and feed providers' bandwidth, by trying to avoid unnecessary feed downloads through the use of conditional HTTP downloading. It saves every feed’s "Last-Modified" and "ETag" response header values (if present) and advises the feed’s HTTP server to only send data if the feed has been updated by modification date/time or "ETag" header. This doesn’t only make feed downloads for RSS feeds with no new updates faster, it also reduces the amount of transferred data per request. Conditional HTTP downloading can be optionally disabled per feed by using the always-download configuration command.

Several aspects of newsboat’s behaviour can be configured via a configuration file config, which is stored next to the urls file. This configuration file contains lines in the form <config-command> <arg1> …​. The configuration file can also contain comments, which start with the # character and go as far as the end of line. If you need to enter a configuration argument that contains spaces, use quotes (") around the whole argument. It’s even possible to integrate the output of external commands into the configuration. The text between two backticks (`) is evaluated as shell command, and its output is put on its place instead. This works like backtick evaluation in Bourne-compatible shells and allows users to use external information from the system within the configuration. Backticks and # characters can be escaped with a backslash (e.g. \` and \#); in that case, they’ll be replaced with literal ` or # in the configuration.

Searching for articles is possible in newsboat, too. Just press the "/" key, enter your search phrase, and the title and content of all articles are searched for it. When you do a search from the list of feeds, all articles of all feeds will be searched. When you do a search from the article list of a feed, only the articles of the currently viewed feed are searched. When opening an article from a search result dialog, the search phrase is highlighted.

The history of all your searches is saved to the filesystem, to the history.search file (stored next to the cache.db file). By default, the last 100 search phrases are stored, but this limited can be influenced through the history-limit configuration variable. To disable search history saving, simply set the history-limit to 0.

3.1. Configuration Commands


Syntax: always-display-description [yes/no]
Default: no
Example: always-display-description yes

If set to yes, then the description will always be displayed even if e.g. a <content:encoded> tag has been found.


Syntax: always-download <url> [<url>…​]
Default: n/a
Example: always-download "https://www.n-tv.de/23.rss"

Specifies one or more feed URLs that should always be downloaded, regardless of their Last-Modified timestamp and ETag header. This option can be specified multiple times.


Syntax: article-sort-order <sortfield>[-<direction>]
Default: date
Example: article-sort-order author-desc

The <sortfield> specifies which article property shall be used for sorting, currently available are: date, title, flags, author, link, guid and random. The optional <direction> specifies the sort direction. asc specifies ascending sorting, desc specifies descending sorting. Note that direction does not affect random sort order. For date, desc is default, for all others, asc is default.


Syntax: articlelist-format <format>
Default: "%4i %f %D %6L %?T?|%-17T| ?%t"
Example: articlelist-format "%4i %f %D %?T?|%-17T| ?%t"

This variable defines the format of entries in the article list. See the respective section in the documentation for more information on format strings.


Syntax: articlelist-title-format <format>
Default: "%N %V - Articles in feed '%T' (%u unread, %t total)%?F? matching filter `%F'&? - %U"
Example: articlelist-title-format "Articles in feed '%T' (%u unread)"

Format of the title in article list. See "Format Strings" section of Newsboat manual for details on available formats.


Syntax: auto-reload [yes/no]
Default: no
Example: auto-reload yes

If set to yes, all feeds will be automatically reloaded at start up and then continuously after a certain time has passed (see reload-time).


Syntax: bind-key <key> <operation> [<dialog>]
Default: n/a
Example: bind-key ^R reload-all

Bind key <key> to <operation>. This means that whenever <key> is pressed, then <operation> is executed (if applicable in the current dialog). A list of available operations can be found below. Optionally, you can specify a dialog. If you specify one, the key binding will only be added to the specified dialog. Available dialogs are all (default if none is specified), feedlist, filebrowser, help, articlelist, article, tagselection, filterselection, urlview, podboat, and dirbrowser.


Syntax: bookmark-autopilot [yes/no]
Default: no
Example: bookmark-autopilot yes

If set to yes, the configured bookmark command is executed without any further input asked from user, unless the url or the title cannot be found/guessed.


Syntax: bookmark-cmd <command>
Default: n/a
Example: bookmark-cmd "~/bin/delicious-bookmark.sh"

If set, then <command> will be used as bookmarking plugin. See the documentation on bookmarking for further information.


Syntax: bookmark-interactive [yes/no]
Default: no
Example: bookmark-interactive yes

If set to yes, then the configured bookmark command is an interactive program.


Syntax: browser <command>
Default: %BROWSER, otherwise lynx
Example: browser "w3m %u"

Set the browser command to use when opening an article in the browser. If the BROWSER environment variable is set, it will be used as the default browser, otherwise lynx will be used. Any occurrences of %u in <command> will be replaced by a URL in single quotes.


Syntax: cache-file <path>
Default: "~/.newsboat/cache.db" or "~/.local/share/cache.db" (see "Files" section)
Example: cache-file "/tmp/testcache.db"

This configuration option sets the cache file. This is especially useful if the filesystem of your home directory doesn’t support proper locking (e.g. NFS).


Syntax: cleanup-on-quit [yes/no]
Default: yes
Example: cleanup-on-quit no

If set to yes, then the cache gets locked and superfluous feeds and items are removed, such as feeds that can’t be found in the urls configuration file anymore.


Syntax: color <element> <fgcolor> <bgcolor> [<attribute> …​]
Default: n/a
Example: color background white black

Set the foreground color, background color and optional attributes for a certain element.


Syntax: confirm-exit [yes/no]
Default: no
Example: confirm-exit yes

If set to yes, then newsboat will ask for confirmation whether the user really wants to quit newsboat.


Set a cookie cache. If set, cookies will be cached in (i.e. read from and written to) this file, using Netscape format.


Syntax: datetime-format <date/time format>
Default: %b %d
Example: datetime-format "%D, %R"

This format specifies the date/time format in the article list. For a detailed documentation on the allowed formats, consult the manpage of strftime(3).


Syntax: define-filter <name> <filterexpr>
Default: n/a
Example: define-filter "all feeds with 'fun' tag" "tags # \"fun\""

With this command, you can predefine filters, which you can later select from a list, and which are then applied after selection. This is especially useful for filters that you need often and you don’t want to enter them every time you need them.


Syntax: delete-read-articles-on-quit [yes/no]
Default: no
Example: delete-read-articles-on-quit yes

If set to yes, then all read articles will be deleted when you quit newsboat.


Syntax: dialogs-title-format <format>
Default: "%N %V - Dialogs"
Example: dialogs-title-format "%N %V - Dialogs"

Format of the title in dialog list. See "Format Strings" section of Newsboat manual for details on available formats.


Syntax: dirbrowser-title-format <format>
Default: "%N %V - %?O?Open Directory&Save File? - %f"
Example: dirbrowser-file-format "%?O?Open Directory&Save File? - %f"

Format of the title in directory browser. See "Format Strings" section of Newsboat manual for details on available formats.


Syntax: display-article-progress [yes/no]
Default: yes
Example: display-article-progress no

If set to yes, then a read progress (in percent) is displayed in the article view. Otherwise, no read progress is displayed.


Syntax: download-full-page [yes/no]
Default: no
Example: download-full-page yes

If set to yes, then for all feed items with no content but with a link, the link is downloaded and the result used as content instead. This may significantly increase the download times of "empty" feeds.


Syntax: download-retries <number>
Default: 1
Example: download-retries 4

How many times newsboat shall try to successfully download a feed before giving up. This is an option to improve the success of downloads on slow and shaky connections such as via a TOR proxy.


Syntax: download-timeout <number>
Default: 30
Example: download-timeout 60

The number of seconds newsboat shall wait when downloading a feed before giving up. This is an option to improve the success of downloads on slow and shaky connections such as via a TOR proxy.


Syntax: error-log <path>
Default: n/a
Example: error-log "~/.newsboat/error.log"

If set, then user errors (e.g. errors regarding defunct RSS feeds) will be logged to this file.


Syntax: external-url-viewer <command>
Default: n/a
Example: external-url-viewer "urlview"

If set, then show-urls will pipe the current article to a specific external tool instead of using the internal URL viewer. This can be used to integrate tools such as urlview.


Syntax: feed-sort-order <sortfield>[-<direction>]
Default: none
Example: feed-sort-order firsttag

The <sortfield> specifies which feed property shall be used for sorting; currently available are: firsttag, title, articlecount, unreadarticlecount, lastupdated and none. The optional <direction> specifies the sort direction. asc specifies ascending sorting, desc specifies descending sorting. desc is the default.


Syntax: feedhq-flag-share <flag>
Default: n/a
Example: feedhq-flag-share "a"

If set and FeedHQ support is used, then all articles that are flagged with the specified flag are being "shared" in FeedHQ so that people that follow you can see it.


Syntax: feedhq-flag-star <flag>
Default: n/a
Example: feedhq-flag-star "b"

If set and FeedHQ support is used, then all articles that are flagged with the specified flag are being "starred" in FeedHQ and appear in the list of "Starred items".


Syntax: feedhq-login <login>
Default: n/a
Example: feedhq-login "your-login"

This variable sets your FeedHQ login for FeedHQ support.


Syntax: feedhq-min-items <number>
Default: 20
Example: feedhq-min-items 100

This variable sets the number of articles that are loaded from FeedHQ per feed.


Syntax: feedhq-password <password>
Default: n/a
Example: feedhq-password "here_goesAquote:\""

This variable sets your FeedHQ password for FeedHQ support. Double quotes should be escaped, i.e. you should write \" instead of ".


Syntax: feedhq-passwordfile <path>
Default: n/a
Example: feedhq-passwordfile "~/.newsboat/feedhq-pw.txt"

A more secure alternative to the above, by storing your password elsewhere in your system.


Syntax: feedhq-passwordeval <command>
Default: n/a
Example: feedhq-passwordeval "gpg --decrypt ~/.newsboat/feedhq-password.gpg"

Another secure alternative, is providing your password from an external command that is evaluated during login. This can be used to read your password from a gpg encrypted file or your system keyring.


Syntax: feedhq-show-special-feeds [yes/no]
Default: yes
Example: feedhq-show-special-feeds "no"

If set and FeedHQ support is used, then "special feeds" like "People you follow" (articles shared by people you follow), "Starred items" (your starred articles) and "Shared items" (your shared articles) appear in your subscription list.


Syntax: feedhq-url <url>
Default: "https://feedhq.org/"
Example: feedhq-url "https://feedhq.example.com/"

Configures the URL where your FeedHQ instance resides.


Syntax: feedlist-format <format>
Default: "%4i %n %11u %t"
Example: feedlist-format " %n %4i - %11u -%> %t"

This variable defines the format of entries in the feed list. See the respective section in the documentation for more information on format strings.


Syntax: feedlist-title-format <format>
Default: "%N %V - %?F?Feeds&Your feeds? (%u unread, %t total)%?F? matching filter `%F'&?%?T? - tag `%T'&?"
Example: feedlist-title-format "Feeds (%u unread, %t total)"

Format of the title in feed list. See "Format Strings" section of Newsboat manual for details on available formats.


Syntax: filebrowser-title-format <format>
Default: "%N %V - %?O?Open File&Save File? - %f"
Example: filebrowser-title-format "%?O?Open File&Save File? - %f"

Format of the title in file browser. See "Format Strings" section of Newsboat manual for details on available formats.


Syntax: goto-first-unread [yes/no]
Default: yes
Example: goto-first-unread no

If set to yes, then the first unread article will be selected whenever a feed is entered.


Syntax: goto-next-feed [yes/no]
Default: yes
Example: goto-next-feed no

If set to yes, then the next-unread, prev-unread and random-unread keys will search in other feeds for unread articles if all articles in the current feed are read. If set to no, then these keys will stop in the current feed.


Syntax: help-title-format <format>
Default: "%N %V - Help"
Example: help-title-format "%N %V - Help"

Format of the title in help window. See "Format Strings" section of Newsboat manual for details on available formats.


Syntax: highlight <target> <regex> <fgcolor> [<bgcolor> [<attribute> …​]]
Default: n/a
Example: highlight all "newsboat" red

With this command, you can highlight text parts in the feed list, the article list and the article view. For a detailed documentation, see the chapter on highlighting.


Syntax: highlight-article <filterexpr> <fgcolor> <bgcolor> [<attribute> …​]
Default: n/a
Example: highlight-article "author =~ \"Andreas Krennmair\"" white red bold

With this command, you can highlight articles in the article list if they match a filter expression. For a detailed documentation, see the chapter on highlighting.


Syntax: history-limit <number>
Default: 100
Example: history-limit 0

Defines the maximum number of entries of commandline resp. search history to be saved. To disable history saving, set it to 0.


Syntax: html-renderer <command>
Default: internal
Example: html-renderer "w3m -dump -T text/html"

If set to internal, then the internal HTML renderer will be used. Otherwise, the specified command will be executed, the HTML to be rendered will be written to the command’s stdin, and the program’s output will be displayed. This makes it possible to use other, external programs, such as w3m, links or lynx, to render HTML.


Syntax: http-auth-method <method>
Default: any
Example: http-auth-method digest

Set HTTP authentication method. Allowed values: any, basic, digest, digest_ie (only available with libcurl 7.19.3 and newer), gssnegotiate, ntlm and anysafe.


Syntax: ignore-article <feed> <filterexpr>
Default: n/a
Example: ignore-article "*" "title =~ \"Windows\""

If a downloaded article from <feed> matches <filterexpr>, then it is ignored and not presented to the user. This command is further explained in the "kill file" section below.


Syntax: ignore-mode [download/display]
Default: download
Example: ignore-mode "display"

This configuration option defines in what way an article is ignored (see ignore-article). If set to download, then it is ignored in the download/parsing phase and thus never written to the cache, if it set to display, it is ignored when displaying articles but is kept in the cache.


Syntax: include <path>
Default: n/a
Example: include "~/.newsboat/colors"

With this command, you can include other files to be interpreted as configuration files. This is especially useful to separate your configuration into several files, e.g. key configuration, color configuration, …​


Syntax: itemview-title-format <format>
Default: "%N %V - Article '%T' (%u unread, %t total)"
Example: itemview-title-format "Article '%T'"

Format of the title in article view. See "Format Strings" section of Newsboat manual for details on available formats.


Syntax: inoreader-app-id <string>
Default: n/a
Example: inoreader-app-id "123456789"

Unique application ID issued by Inoreader. See "Inoreader" section.


Syntax: inoreader-app-key <string>
Default: n/a
Example: inoreader-app-key "TmV3c2JvYXQgcm9ja3MgOikK"

Application key issued by Inoreader. See "Inoreader" section.


Syntax: inoreader-flag-share <flag>
Default: n/a
Example: inoreader-flag-share "a"

If set and Inoreader support is used, then all articles that are flagged with the specified flag are being "shared" in Inoreader so that people that follow you can see it.


Syntax: inoreader-flag-star <flag>
Default: n/a
Example: inoreader-flag-star "b"

If set and Inoreader support is used, then all articles that are flagged with the specified flag are being "starred" in Inoreader and appear in the list of "Starred items".


Syntax: inoreader-login <login>
Default: n/a
Example: inoreader-login "your-login"

This variable sets your Inoreader login for Inoreader support.


Syntax: inoreader-min-items <number>
Default: 20
Example: inoreader-min-items 100

This variable sets the number of articles that are loaded from Inoreader per feed.


Syntax: inoreader-password <password>
Default: n/a
Example: inoreader-password "here_goesAquote:\""

This variable sets your Inoreader password for Inoreader support. Double quotes should be escaped, i.e. you should write \" instead of ".


Syntax: inoreader-passwordfile <path>
Default: n/a
Example: inoreader-passwordfile "~/.newsboat/inoreader-pw.txt"

A more secure alternative to the above, by storing your password elsewhere in your system.


Syntax: inoreader-passwordeval <command>
Default: n/a
Example: inoreader-passwordeval "gpg --decrypt ~/.newsboat/inoreader-password.gpg"

Another secure alternative, is providing your password from an external command that is evaluated during login. This can be used to read your password from a gpg encrypted file or your system keyring.


Syntax: inoreader-show-special-feeds [yes/no]
Default: yes
Example: inoreader-show-special-feeds "no"

If set and Inoreader support is used, then "special feeds" like "Starred items" (your starred articles) and "Shared items" (your shared articles) appear in your subscription list.


Syntax: keep-articles-days <number>
Default: 0
Example: keep-articles-days 30

If set to a number greater than 0, only articles that were published within the last <number> days are kept, and older articles are deleted. If set to 0, this option is not active. Note that changing this setting won’t bring back the articles that were deleted earlier; currently, there’s no non-hacky way to bring back deleted articles.


Syntax: macro <macro key> <command list>
Default: n/a
Example: macro k open; reload; quit

With this command, you can define a macro key and specify a list of commands that shall be executed when the macro prefix and the macro key are pressed.


Syntax: mark-as-read-on-hover [yes/no]
Default: no
Example: mark-as-read-on-hover yes

If set to yes, then all articles that get selected in the article list are marked as read.


Syntax: max-download-speed <number>
Default: 0
Example: max-download-speed 50

If set to a number greater than 0, the download speed per download is set to that limit (in KB/s).


Syntax: max-browser-tabs <number>
Default: 10
Example: max-browser-tabs 4

Set the maximum number of articles to open in a browser when using the open-all-unread-in-browser or open-all-unread-in-browser-and-mark-read commands.


Syntax: max-items <number>
Default: 0
Example: max-items 100

Set the number of articles to maximally keep per feed. If the number is set to 0, then all articles are kept.


Syntax: miniflux-login <username>
Default: n/a
Example: miniflux-login "admin"

Sets the username for use with Miniflux.


Syntax: miniflux-password <password>
Default: n/a
Example: miniflux-password "here_goesAquote:\""

Configures the password for use with Miniflux. Double quotes should be escaped, i.e. you should write \" instead of ".


Syntax: miniflux-passwordfile <path>
Default: n/a
Example: miniflux-passwordfile "~/.newsboat/miniflux-pw.txt"

A more secure alternative to the above, by storing your password elsewhere in your system.


Syntax: miniflux-passwordeval <command>
Default: n/a
Example: miniflux-passwordeval "gpg --decrypt ~/.newsboat/miniflux-password.gpg"

Another secure alternative, is providing your password from an external command that is evaluated during login. This can be used to read your password from a gpg encrypted file or your system keyring.


Syntax: miniflux-url <url>
Default: n/a
Example: miniflux-url "https://example.com/miniflux/"

Configures the URL where the Miniflux installation you want to use resides.


Syntax: newsblur-login <login>
Default: n/a
Example: newsblur-login "your-login"

This variable sets your NewsBlur login for NewsBlur support.


Syntax: newsblur-min-items <number>
Default: 20
Example: newsblur-min-items 100

This variable sets the number of articles that are loaded from NewsBlur per feed.


Syntax: newsblur-password <password>
Default: n/a
Example: newsblur-password "here_goesAquote:\""

This variable sets your NewsBlur password for NewsBlur support. Double quotes should be escaped, i.e. you should write \" instead of ".


Syntax: newsblur-passwordfile <path>
Default: n/a
Example: newsblur-passwordfile "~/.newsboat/newsblur-pw.txt"

A more secure alternative to the above, by storing your password elsewhere in your system.


Syntax: newsblur-passwordeval <command>
Default: n/a
Example: newsblur-passwordeval "gpg --decrypt ~/.newsboat/newsblur-password.gpg"

Another secure alternative, is providing your password from an external command that is evaluated during login. This can be used to read your password from a gpg encrypted file or your system keyring.


Syntax: newsblur-url <url>
Default: "https://newsblur.com"
Example: newsblur-url "https://localhost"

Configures the URL where the NewsBlur instance resides.


Syntax: notify-always [yes/no]
Default: no
Example: notify-always yes

If set to no, notifications will only be made when there are new feeds or articles. If set to yes, notifications will be made regardless.


Syntax: notify-beep [yes/no]
Default: no
Example: notify-beep yes

If set to yes, then the speaker will beep on new articles.


Syntax: notify-format <string>
Default: "newsboat: finished reload, %f unread feeds (%n unread articles total)"
Example: notify-format "%d new articles (%n unread articles, %f unread feeds)"

Format string that is used for formatting notifications. See the chapter on format strings for more information.


Syntax: notify-program <command>
Default: n/a
Example: notify-program "~/bin/my-notifier"

If set, then the configured program will be executed if new articles arrived (through a reload) or if notify-always is yes. The first parameter of the called program contains the notification message. In order to pass other hard-coded arguments to the program, write an appropriate wrapper shell script and use it as <command> instead.


Syntax: notify-screen [yes/no]
Default: no
Example: notify-screen yes

If set to yes, then a "privacy message" will be sent to the terminal, containing a notification message about new articles. This is especially useful if you use terminal emulations such as GNU screen which implement privacy messages.


Syntax: notify-xterm [yes/no]
Default: no
Example: notify-xterm yes

If set to yes, then the xterm window title will be set to a notification message about new articles.


Syntax: ocnews-flag-star <character>
Default: n/a
Example: ocnews-flag-star "s"

If set and ownCloud News support is used, then all articles that are flagged with the specified flag are being "starred" in ownCloud News.


Syntax: ocnews-login <username>
Default: n/a
Example: ocnews-login "user"

Sets the username to use with the ownCloud instance.


Syntax: ocnews-password <password>
Default: n/a
Example: ocnews-password "here_goesAquote:\""

Configures the password to use with the ownCloud instance. Double quotes should be escaped, i.e. you should write \" instead of ".


Syntax: ocnews-passwordfile <path>
Default: n/a
Example: ocnews-passwordfile "~/.newsboat/ocnews-pw.txt"

A more secure alternative to the above, by storing your password elsewhere in your system.


Syntax: ocnews-passwordeval <command>
Default: n/a
Example: ocnews-passwordeval "gpg --decrypt ~/.newsboat/ocnews-password.gpg"

Another secure alternative, is providing your password from an external command that is evaluated during login. This can be used to read your password from a gpg encrypted file or your system keyring.


Syntax: ocnews-url <url>
Default: n/a
Example: ocnews-url "https://localhost/owncloud"

Configures the URL where the ownCloud instance resides.


Syntax: oldreader-flag-share <flag>
Default: n/a
Example: oldreader-flag-share "a"

If set and The Old Reader support is used, then all articles that are flagged with the specified flag are being "shared" in The Old Reader so that people that follow you can see it.


Syntax: oldreader-flag-star <flag>
Default: n/a
Example: oldreader-flag-star "b"

If set and The Old Reader support is used, then all articles that are flagged with the specified flag are being "starred" in The Old Reader and appear in the list of "Starred items".


Syntax: oldreader-login <login>
Default: n/a
Example: oldreader-login "your-login"

This variable sets your The Old Reader login for The Older Reader support.


Syntax: oldreader-min-items <number>
Default: 20
Example: oldreader-min-items 100

This variable sets the number of articles that are loaded from The Old Reader per feed.


Syntax: oldreader-password <password>
Default: n/a
Example: oldreader-password "here_goesAquote:\""

This variable sets your The Old Reader password for The Old Reader support. Double quotes should be escaped, i.e. you should write \" instead of ".


Syntax: oldreader-passwordfile <path>
Default: n/a
Example: oldreader-passwordfile "~/.newsboat/oldreader-pw.txt"

A more secure alternative to the above, by storing your password elsewhere in your system.


Syntax: oldreader-passwordeval <command>
Default: n/a
Example: oldreader-passwordeval "gpg --decrypt ~/.newsboat/oldreader-password.gpg"

Another secure alternative, is providing your password from an external command that is evaluated during login. This can be used to read your password from a gpg encrypted file or your system keyring.


Syntax: oldreader-show-special-feeds [yes/no]
Default: yes
Example: oldreader-show-special-feeds "no"

If set and The Old reader support is used, then "special feeds" like "People you follow" (articles shared by people you follow), "Starred items" (your starred articles) and "Shared items" (your shared articles) appear in your subscription list.


Syntax: openbrowser-and-mark-jumps-to-next-unread [yes/no]
Default: no
Example: openbrowser-and-mark-jumps-to-next-unread yes

If set to yes, jump to the next unread item when an item is opened in the browser and marked as read.


Syntax: opml-url <url> …​
Default: n/a
Example: opml-url "https://host.domain.tld/blogroll.opml" "https://example.com/anotheropmlfile.opml"

If the OPML online subscription mode is enabled, then the list of feeds will be taken from the OPML file found on this location. Optionally, you can specify more than one URL. All the listed OPML URLs will then be taken into account when loading the feed list.


Syntax: pager [<command>/internal]
Default: internal
Example: pager "less %f"

If set to internal, then the internal pager will be used. Otherwise, the article to be displayed will be rendered to be a temporary file and then displayed with the configured pager. If the command is set to an empty string, the content of the PAGER environment variable will be used. If the command contains a placeholder %f, it will be replaced with the temporary filename.


Syntax: podcast-auto-enqueue [yes/no]
Default: no
Example: podcast-auto-enqueue yes

If set to yes, then all podcast URLs that are found in articles are added to the podcast download queue. See the respective section in the documentation for more information on podcast support in newsboat.


Syntax: prepopulate-query-feeds [yes/no]
Default: no
Example: prepopulate-query-feeds yes

If set to yes, then all query feeds are prepopulated with articles on startup.


Syntax: ssl-verifyhost [yes/no]
Default: yes
Example: ssl-verifyhost no

If set to no, skip verification of the certificate’s name against host.


Syntax: ssl-verifypeer [yes/no]
Default: yes
Example: ssl-verifypeer no

If set to no, skip verification of the peer’s SSL certificate.


Syntax: proxy-auth-method <method>
Default: any
Example: proxy-auth-method ntlm

Set proxy authentication method. Allowed values: any, basic, digest, digest_ie (only available with libcurl 7.19.3 and newer), gssnegotiate, ntlm and anysafe.


Syntax: proxy-auth <auth>
Default: n/a
Example: proxy-auth user:password

Set the proxy authentication string.


Syntax: proxy-type <type>
Default: http
Example: proxy-type socks5

Set proxy type. Allowed values: http, socks4, socks4a, socks5 and socks5h.


Syntax: proxy <server:port>
Default: n/a
Example: proxy localhost:3128

Set the proxy to use for downloading RSS feeds. (Don’t forget to actually enable the proxy with use-proxy yes.)


Syntax: refresh-on-startup [yes/no]
Default: no
Example: refresh-on-startup yes

If set to yes, then all feeds will be reloaded when newsboat starts up. This is equivalent to the -r commandline option.


Syntax: reload-only-visible-feeds [yes/no]
Default: no
Example: reload-only-visible-feeds yes

If set to yes, then manually reloading all feeds will only reload the currently visible feeds, e.g. if a filter or a tag is set.


Syntax: reload-threads <number>
Default: 1
Example: reload-threads 3

The number of parallel reload threads that shall be started when all feeds are reloaded.


Syntax: reload-time <number>
Default: 60
Example: reload-time 120

The number of minutes between automatic reloads.


Syntax: reset-unread-on-update <url> [<url>…​]
Default: n/a
Example: reset-unread-on-update "https://blog.fefe.de/rss.xml?html"

Specifies one or more feed URLs for whose articles the unread flag will be reset if an article has been updated, i.e. its content has been changed. This is especially useful for RSS feeds where single articles are updated after publication, and you want to be notified of the updates. This option can be specified multiple times.


Syntax: run-on-startup <list of operations>
Default: n/a
Example: run-on-startup next-unread; open; random-unread; open

Specifies one or more operations, separated by semicolons, which are executed on Newsboat startup.


Syntax: save-path <path-to-directory>
Default: ~/
Example: save-path "~/Saved Articles"

The default path where articles shall be saved to. If an invalid path is specified, the current directory is used.


Syntax: scrolloff <number>
Default: 0
Example: scrolloff 5

Keep the configured number of lines above and below the selected item in lists. Configure a high number to keep the selected item in the center of the screen.


Syntax: search-highlight-colors <fgcolor> <bgcolor> [<attribute> …​]
Default: black yellow bold
Example: search-highlight-colors white black bold

This configuration command specifies the highlighting colors when searching for text from the article view.


Syntax: searchresult-title-format <format>
Default: "%N %V - Search results (%u unread, %t total)%?F? matching filter `%F'&?"
Example: searchresult-title-format "Search result"

Format of the title in search result. See "Format Strings" section of Newsboat manual for details on available formats.


Syntax: selectfilter-title-format <format>
Default: "%N %V - Select Filter"
Example: selectfilter-title-format "Select Filter"

Format of the title in filter selection dialog. See "Format Strings" section of Newsboat manual for details on available formats.


Syntax: selecttag-format <format>
Default: "%4i %T (%u)"
Example: selecttag-format "[%2i] %T (%n unread articles in %f feeds, %u feeds total)"

Format of the lines in "Select tag" dialog. See the respective section in the documentation for more information on format strings.


Syntax: selecttag-title-format <format>
Default: "%N %V - Select Tag"
Example: selecttag-title-format "Select Tag"

Format of the title in tag selection dialog. See "Format Strings" section of Newsboat manual for details on available formats.


Syntax: show-keymap-hint [yes/no]
Default: yes
Example: show-keymap-hint no

If set to no, then the keymap hints on the bottom of screen will not be displayed.


Syntax: show-title-bar [yes/no]
Default: yes
Example: show-title-bar no

If set to no, then the title bar on the top of the screen will not be displayed.


Syntax: show-read-articles [yes/no]
Default: yes
Example: show-read-articles no

If set to yes, then all articles of a feed are listed in the article list. If set to no, then only unread articles are listed.


Syntax: show-read-feeds [yes/no]
Default: yes
Example: show-read-feeds no

If set to yes, then all feeds, including those without unread articles, are listed. If set to no, then only feeds with one or more unread articles are list.


Syntax: suppress-first-reload [yes/no]
Default: no
Example: suppress-first-reload yes

If set to yes, then the first automatic reload will be suppressed if auto-reload is set to yes.


Syntax: swap-title-and-hints [yes/no]
Default: no
Example: swap-title-and-hints yes

If set to yes, then the title at the top of screen and keymap hints at the bottom of screen will be swapped.


Syntax: text-width <number>
Default: 0
Example: text-width 72

If set to a number greater than 0, all HTML will be rendered to this maximum line length or the terminal width (whichever is smaller). If set to 0, the terminal width will always be used. Does not apply when using external renderer or viewing the source. Also note that "Link" header and "Links" section won’t be affected by it—they contain URLs which are better not wrapped.


Syntax: toggleitemread-jumps-to-next-unread [yes/no]
Default: no
Example: toggleitemread-jumps-to-next-unread yes

If set to yes, jump to the next unread item when an item’s read status is toggled in the article list.


Syntax: ttrss-flag-publish <character>
Default: n/a
Example: ttrss-flag-publish "b"

If set and Tiny Tiny RSS support is used, then all articles that are flagged with the specified flag are being marked as "published" in Tiny Tiny RSS.


Syntax: ttrss-flag-star <character>
Default: n/a
Example: ttrss-flag-star "a"

If set and Tiny Tiny RSS support is used, then all articles that are flagged with the specified flag are being "starred" in Tiny Tiny RSS.


Syntax: ttrss-login <username>
Default: n/a
Example: ttrss-login "admin"

Sets the username for use with Tiny Tiny RSS.


Syntax: ttrss-mode [multi/single]
Default: multi
Example: ttrss-mode "single"

Configures the mode in which Tiny Tiny RSS is used. In single-user mode, login and password are used for HTTP authentication, while in multi-user mode, they are used for authenticating with Tiny Tiny RSS.


Syntax: ttrss-password <password>
Default: n/a
Example: ttrss-password "here_goesAquote:\""

Configures the password for use with Tiny Tiny RSS. Double quotes should be escaped, i.e. you should write \" instead of ".


Syntax: ttrss-passwordfile <path>
Default: n/a
Example: ttrss-passwordfile "~/.newsboat/ttrss-pw.txt"

A more secure alternative to the above, by storing your password elsewhere in your system.


Syntax: ttrss-passwordeval <command>
Default: n/a
Example: ttrss-passwordeval "gpg --decrypt ~/.newsboat/ttrss-password.gpg"

Another secure alternative, is providing your password from an external command that is evaluated during login. This can be used to read your password from a gpg encrypted file or your system keyring.


Syntax: ttrss-url <url>
Default: n/a
Example: ttrss-url "https://example.com/ttrss/"

Configures the URL where the Tiny Tiny RSS installation you want to use resides.


Syntax: unbind-key <key> [<dialog>]
Default: n/a
Example: unbind-key R

Unbind key <key>. This means that no operation is called when <key> is pressed. If you provide "-a" as <key>, all currently bound keys will become unbound. Optionally, you can specify a dialog (for a list of available dialogs, see bind-key above). If you specify one, the key binding will only be unbound for the specified dialog.


Syntax: urls-source <source>
Default: "local"
Example: urls-source "oldreader"

This configuration command sets the source where URLs shall be retrieved from. By default, this is the urls file. Alternatively, you can set it to opml, which enables newsboat’s OPML online subscription mode, to ttrss which enables newsboat’s Tiny Tiny RSS support, to oldreader, which enables newsboat’s The Old Reader support, to newsblur, which enables NewsBlur support, to feedhq for FeedHQ support, to ocnews for ownCloud News support, to inoreader for Inoreader support, or to miniflux for Miniflux support. Query feed specifications will be read from the local urls file regardless of this setting.


Syntax: urlview-title-format <format>
Default: "%N %V - URLs"
Example: urlview-title-format "URLs"

Format of the title in URL view. See "Format Strings" section of Newsboat manual for details on available formats.


Syntax: use-proxy [yes/no]
Default: no
Example: use-proxy yes

If set to yes, then the configured proxy will be used for downloading the RSS feeds.


Syntax: user-agent <string>
Default: n/a
Example: user-agent "Lynx/2.8.5rel.1 libwww-FM/2.14"

If set to a non-zero-length string, this value will be used as HTTP User-Agent header for all HTTP requests.


Syntax: wrap-scroll [yes/no]
Default: no
Example: wrap-scroll yes

If set to yes, moving down while on the last item in a list will wrap around to the top and vice versa.

3.2. Available Operations


Operation: open
Default key: ENTER

Open the currently selected feed or article.


Operation: quit
Default key: q

Quit the program or return to the previous dialog (depending on the context).


Operation: hard-quit
Default key: Q

Quit the program without confirmation.


Operation: reload
Default key: r

Reload the currently selected feed.


Operation: reload-all
Default key: R

Reload all feeds.


Operation: mark-feed-read
Default key: A

Mark all articles in the currently selected feed read.


Operation: mark-all-feeds-read
Default key: C

Mark articles in all feeds read.


Operation: mark-all-above-as-read
Default key: n/a

Mark all above as read.


Operation: save
Default key: s

Save the currently selected article to a file.


Operation: save-all
Default key: n/a

Save all articles from currently selected feed.


Operation: next-unread
Default key: n

Jump to the next unread article.


Operation: prev-unread
Default key: p

Jump to the previous unread article.


Jump to next list entry.


Jump to previous list entry.


Operation: random-unread
Default key: ^K

Jump to a random unread article.


Operation: open-in-browser
Default key: o

Open the URL associated with the current article, or selection when in the URL view.


Operation: open-in-browser-and-mark-read
Default key: O

Open the URL associated with the current article, or selection when in the URL view. When used in the article view, it will also mark the article as read.


Operation: open-all-unread-in-browser
Default key: n/a

Open all the unread URLs in the current feed.


Operation: open-all-unread-in-browser-and-mark-read
Default key: n/a

Open all the unread URLs in the current feed and mark them as read.


Operation: help
Default key: ?

Run the help screen.


Operation: toggle-source-view
Default key: ^U

Toggle between the HTML view and the source view in the article view.


Operation: toggle-article-read
Default key: N

Toggle the read flag for the currently selected article, and clear the delete flag if set.


Operation: toggle-show-read-feeds
Default key: l

Toggle whether read feeds should be shown in the feed list.


Operation: show-urls
Default key: u

Show all URLs in the article in a list (similar to urlview).


Operation: clear-tag
Default key: ^T

Clear current tag.


Operation: set-tag
Default key: t

Select tag.


Open the search dialog. When a search is done in the article list, then the search operation only applies to the articles of the current feed, otherwise to all articles.


Operation: goto-url
Default key: #

Open the URL dialog and then open a specified URL in the browser.


Operation: one
Default key: 1

Open URL 1 in the browser.


Operation: two
Default key: 2

Open URL 2 in the browser.


Operation: three
Default key: 3

Open URL 3 in the browser.


Operation: four
Default key: 4

Open URL 4 in the browser.


Operation: five
Default key: 5

Open URL 5 in the browser.


Operation: six
Default key: 6

Open URL 6 in the browser.


Operation: seven
Default key: 7

Open URL 7 in the browser.


Operation: eight
Default key: 8

Open URL 8 in the browser.


Operation: nine
Default key: 9

Open URL 9 in the browser.


Operation: zero
Default key: 0

Open URL 10 in the browser.


Operation: enqueue
Default key: e

Add the podcast download URL of the current article (if any is found) to the podcast download queue (see the respective section in the documentation for more information on podcast support).


Operation: edit-urls
Default key: E

Edit the list of subscribed URLs. newsboat will start the editor configured through the VISUAL environment variable (if unset, EDITOR is used; fallback: vi). When editing is finished, newsboat will reload the URLs file.


Operation: reload-urls
Default key: ^R

Reload the URLs configuration file.


Operation: redraw
Default key: ^L

Redraw the screen.


Operation: cmdline
Default key: :

Open the command line.


Operation: set-filter
Default key: F

Set a filter.


Operation: select-filter
Default key: f

Select a predefined filter.


Operation: clear-filter
Default key: ^F

Clear currently set filter.


Operation: bookmark
Default key: ^B

Bookmark currently selected article or URL.


Operation: edit-flags
Default key: ^E

Edit the flags of the currently selected article.


Operation: next-unread-feed
Default key: ^N

Go to the next feed with unread articles. This only works from the article list.


Operation: prev-unread-feed
Default key: ^P

Go to the previous feed with unread articles. This only works from the article list.


Operation: next-feed
Default key: j

Go to the next feed. This only works from the article list.


Operation: prev-feed
Default key: k

Go to the previous feed. This only works from the article list.


Operation: delete-article
Default key: D

Delete the currently selected article.


Operation: delete-all-articles
Default key: ^D

Delete all articles in the current feed.


Operation: purge-deleted
Default key: $

Purge all articles that are marked as deleted from the article list.


Operation: view-dialogs
Default key: v

View list of open dialogs.


Operation: close-dialog
Default key: ^X

Close currently selected dialog.


Operation: next-dialog
Default key: ^V

Go to next dialog.


Operation: prev-dialog
Default key: ^G

Go to previous dialog.


Operation: pipe-to
Default key: |

Pipe article to command.


Operation: sort
Default key: g

Sort feeds/articles by interactively choosing the sort method.


Operation: rev-sort
Default key: G

Sort feeds/articles by interactively choosing the sort method (reversed).


Operation: up
Default key: UP

Go up one item in the list.


Operation: down
Default key: DOWN

Go down one item in the list.


Operation: pageup
Default key: PPAGE

Go up one page in the list.


Operation: pagedown
Default key: NPAGE

Go down one page in the list.


Operation: home
Default key: HOME

Go to the first item in the list.


Operation: end
Default key: END

Go to the last item in the list.


Operation: macro-prefix
Default key: ,

Initiate macro execution. The next key press selects the actual macro and runs it.


Operation: switch-focus
Default key: TAB

Switch focus between widgets. This is currently only applicable to the filebrowser and dirbrowser contexts.

Keys, as used in the bind-key configuration command, use a special syntax. Lowercase keys, uppercase keys and special characters are written literally. The "Enter" key is written as ENTER, "Tab" key as TAB, and the "Esc" key is written as ESC. The function keys "F1" to "F12" are written as F1 to F12. The "Space" key is written as SPACE. Key combinations with the "Ctrl" key, such as "Ctrl-R", are written as ^R. Please be aware that all Ctrl-related key combinations need to be written in uppercase. The following identifiers for keys are supported:

  • ENTER (Enter key)

  • BACKSPACE (backspace key)

  • LEFT (left cursor)

  • RIGHT (right cursor)

  • UP (up cursor)

  • DOWN (down cursor)

  • PPAGE (page up cursor)

  • NPAGE (page down cursor)

  • HOME (cursor to beginning of list/article)

  • END (cursor to end of list/article)

  • ESC (Esc key)

  • TAB (Tab key)

3.3. Example Configuration

# a comment
max-items        100 # such comments are possible, too
browser          links
show-read-feeds  no

unbind-key       R
bind-key         ^R    reload-all

3.4. Using Double Quotes

TL;DR: use double quotes for strings that contain spaces or double quotes. Escape double quotes (use \") and backslashes (use \\). Don’t escape stuff outside of double quotes, and don’t use single quotes for quoting — Newsboat doesn’t support that.

Many of Newsboat’s options expect strings as arguments, be it commands, passwords, dialog titles, URLs etc. Some options even take multiple strings at once. These strings can contain spaces, which might confuse Newsboat since it already uses spaces to separate option names from option arguments.

To help Newsboat understand your intent, put such strings into double quotes:

browser "firefox --new-tab %u"

What if you need a double quote inside a string? Escape it with a backslash:

ocnews-password "UnbalancedQuotes\"AreSoFun!"

And what about the backslash itself? Escape it, too! Suppose you have a program called my favourite pager, and you want to view articles with it. Newsboat ultimately passes commands to the shell, and shell expects spaces to be escaped if you want them preserved. But since Newsboat interprets backslashes, you have to add another layer of escaping. Thus, you end up with a command like this:

pager "/usr/bin/my\\ favourite\\ pager"

3.5. Configuring Colors

It is possible to configure custom color settings in newsboat. The basic configuration syntax is:

color <element> <foreground color> <background color> [<attribute> ...]

This means that if you configure colors for a certain element, you need to provide a foreground color and a background color as a minimum. The following colors are supported:

  • black

  • red

  • green

  • yellow

  • blue

  • magenta

  • cyan

  • white

  • default

  • color<n>, e.g. color123

The default color means that the terminal’s default color will be used. The color<n> color name (where <n> is a decimal number not starting with zero) can be used if your terminal support 256 colors (e.g. gnome-terminal, xterm with $TERM set to xterm-256color). Newsboat contains support for 256 color terminals since version 2.1. For a complete chart of colors and their corresponding numbers, please see https://www.calmar.ws/vim/256-xterm-24bit-rgb-color-chart.html.

Optionally, you can also add one or more attributes. The following attributes are supported:

  • standout

  • underline

  • reverse

  • blink

  • dim

  • bold

  • protect

  • invis

Currently, the following elements are supported:

  • listnormal: a normal list item

  • listfocus: the currently selected list item

  • listnormal_unread: an unread list item

  • listfocus_unread: the currently selected and unread list item

  • info: the info bars on top and bottom

  • background: the application background

  • article: the article text

  • end-of-text-marker: filler lines (~) below blocks of text

The default color configuration of newsboat looks like this:

color background          white   black
color listnormal          white   black
color listfocus           yellow  blue   bold
color listnormal_unread   magenta black
color listfocus_unread    magenta blue   bold
color info                yellow  blue   bold
color article             white   black

3.6. Files

By default, Newsboat stores all the files in a traditional Unix fashion, i.e. in the "dotdir" located at ~/.newsboat. However, it also supports a modern way, XDG Base Directory Specification, which splits the files between the following locations:

  • ~/.local/share/newsboat/

  • ~/.config/newsboat/

If these directories exist or the environment variables $XDG_CONFIG_HOME and $XDG_DATA_HOME are set, newsboat will use these directories, otherwise it will default to ~/.newsboat as its configuration directory.

If you’re currently using ~/.newsboat/ but wish to migrate to XDG directories, you should move the files as follows:

config, urls

to $HOME/.config/newsboat/

cache.db, history.search, history.cmdline, queue

to $HOME/.local/share/newsboat/

3.7. Migrating from other RSS Feed Readers

It is very likely that you have used other RSS feed readers before. In this case, it is practical to migrate the previous configuration to newsboat.

3.7.1. Newsbeuter (automatic migration)

Newsboat is a fork of Newsbeuter, so the migration from the latter is partially automated. Simple enough configurations will be transferred without any user intervention, while more complicated ones might require a little tweaking afterwards.

To prevent data loss, please check the results before deleting your old configuration. Pay extra attention to files that you include in your configuration—​you’d probably want to manually copy them over to Newsboat directories, and possibly update the include paths.

Migration from Newsbeuter is attempted on startup if the following conditions are met:

  • none of -u, -c, -C options were specified; and

  • the urls file doesn’t exist (neither ~/.newsboat/urls nor $XDG_CONFIG_HOME/newsboat/urls).

Newsboat first tries to migrate an XDG configuration, and if that fails, it tries the dotdir one. If that fails as well, Newsboat proceeds with the startup as usual.

XDG migration checks that:

  • $XDG_CONFIG_HOME/newsbeuter/ is readable and executable; and

  • neither $XDG_CONFIG_HOME/newsboat/ nor $XDG_DATA_HOME/newsboat/ exist.

If both conditions are met, Newsboat tries to create its XDG dirs (aborting the migration if that fails), then copies the following files: urls and config from config dir; cache.db, queue, history.search, history.cmdline from data dir.

Dotdir migration checks that ~/.newsbeuter/ is readable and executable, and ~/.newsboat/ doesn’t exist. If those conditions are met, it tries to create ~/.newsboat/ (aborting the migration if that fails), then copies the following files: urls, cache.db, config, queue, history.search, history.cmdline.

There’s one scenario where the process breaks: if you have an XDG configuration for Newsbeuter, ~/.newsboat/ exists and ~/.newsboat/urls doesn’t exist, then Newsboat will migrate XDG files and proceed, ignoring the dotdir. Please don’t confuse the poor program like that!

3.7.2. Other readers (via OPML)

The vast amount of RSS feed readers allows the export of subscriptions via OPML files. OPML is an XML file format that was designed to save outlines, and has found its primary use in the import and export of feed subscriptions between different RSS feed readers.

The best thing to start with is to export your subscriptions from the old reader. Usually, RSS feed readers have appropriate menu items available to do so.

Snownews provides a script to convert your current subscription file into an OPML file:

snow2opml > ~/blogroll.opml

This command creates from your Snownews configuration a file blogroll.opml in your home directory. To export the subscription list from raggle, the following command is necessary:

raggle --export-opml ~/blogroll.opml

When you have exported the subscriptions from your old RSS feed reader, you can import them into newsboat:

newsboat -i ~/blogroll.opml

Don’t worry, newsboat won’t destroy your existing configuration, or add subscriptions more than once: every URL that is added to the subscription list is checked before whether it is already in the list, and is only added if not. This makes it possible to merge several OPML files into your subscription list.

If your old RSS feed reader was able to structure your subscriptions in hierarchies, and reflected this structure in the exported OPML file, newsboat doesn’t throw away this information (although it doesn’t support hierarchies), but generates tags from it. Tags are newsboat’s way of organizing subscriptions in a non-hierarchical way. More information on the use of tags can be found below.

Imagine the following folder hierarchy:

|- News
| |- Europe
| `- International
|- IT
| |- Linux
| |- Windows
| `- Programming
|   |- C++
|   |- Ruby
|   `- Erlang
`- Private

Subscriptions found in the folder "Private" will be tagged with "Private", subscriptions in the folder "International" will be tagged with "News" and "News/International", subscriptions in the folder "Erlang" will be tagged ith "IT", "IT/Programming" and "IT/Programming/Erlang", and so on. This means that when you select the tag "Programming" in newsboat, you will see all subscriptions that were in the "Programming" folder or one of its subfolders before. This means that you will lose virtually nothing of your previously configured structure.

4. Newsboat as a Client for Newsreading Services

Newsboat is a perfectly capable feedreader in its own right, but it can’t cover all the use-cases. For example, you might want to synchronize your news between devices, or you don’t have a computer that’s running 24/7 to fetch your feeds. In such cases, you might register with some online feedreader, and use Newsboat as a client. The next few sections provide configuration instructions for each supported service.

4.1. The Old Reader

The Old Reader is a successor to Google Reader. In order to use it, you first need to configure the proper URL source:

urls-source "oldreader"

In addition, Newsboat needs to know your The Old Reader username and password so that it can authenticate with the service:

oldreader-login "your-oldreader-account"
oldreader-password "your-password"

Note that double quotes should be escaped, i.e. you should write \" instead of ".

After setting these configuration values, you can start Newsboat. It will authenticate with The Old Reader and download your subscription list.

When you mark single items or complete feeds as read, Newsboat will synchronize this information directly to The Old Reader. This, of course, includes opening articles. Toggling read articles back to "unread" is also communicated to The Old Reader.

In addition, The Old Reader provides the ability to "star" and to "share" articles. Starred articles are basically bookmarks, while shared articles are shown to people that follow your The Old Reader account. Newsboat allows the use of this feature by mapping its powerful "flags" to the "star"/"unstar" resp. "share"/"unshare" operations.

In order to use this mapping, all you need to do is to configure the flags that shall be used:

oldreader-flag-share "a"
oldreader-flag-star "b"

After that, use these flags when you edit flags for an article, and these articles will be starred resp. shared.

By default, Newsboat also shows The Old Reader "special feeds":

  • People you follow: articles shared by people that you follow.

  • Starred items: articles that you starred.

  • Shared items: articles that you shared.

You can disable these feeds by setting the following configuration variable:

oldreader-show-special-feeds no

The Old Reader’s folders are converted into Newsboat tags. You can select and filter feeds by tags; see Tagging and Filter Language for details.

4.2. NewsBlur

NewsBlur is a successor to Google Reader. To use it, set your urls-source:

urls-source "newsblur"

Then, configure your NewsBlur credentials:

newsblur-login "your-newsblur-account"
newsblur-password "your-password"

Note that double quotes should be escaped, i.e. you should write \" instead of ".

Finally, set a path to the file where Newsboat can store the HTTP cookies:

cookie-cache "~/.newsboat/cookies.txt"

When you start Newsboat, it will download the feeds that you configured in NewsBlur. Please take a closer look at the configuration commands for what you can configure in Newsboat regarding NewsBlur.

NewsBlur’s folders are converted into Newsboat tags. You can select and filter feeds by tags; see Tagging and Filter Language for details.

4.3. FeedHQ

FeedHQ is a successor to Google Reader. Configuration basically works the same as with The Old Reader.

First, set your urls-source:

urls-source "feedhq"

Then, configure your FeedHQ credentials:

feedhq-login "your-feedhq-account"
feedhq-password "your-password"

Note that double quotes should be escaped, i.e. you should write \" instead of ".

If you’re using a standalone instance, you should add one more setting:

feedhq-url "https://the.url.of/your/feedhq/instance"

When you start Newsboat, it will download the feeds that you configured in FeedHQ. Please take a closer look at the configuration commands for what you can configure in Newsboat regarding FeedHQ.

FeedHQ’s folders are converted into Newsboat tags. You can select and filter feeds by tags; see Tagging and Filter Language for details.

4.4. Bazqux

Bazqux uses "Google Reader API", which is also used by other readers like FeedHQ. Thus, one can leverage Newsboat’s support for FeedHQ (described above) to synchronize with Bazqux:

urls-source "feedhq"
feedhq-url "https://bazqux.com/"
feedhq-login "username"
feedhq-password "password"

See Passwords for external APIs for other, more secure ways to store your password.

4.5. Tiny Tiny RSS

Newsboat can be used to synchronize with Tiny Tiny RSS installations. Tiny Tiny RSS is a web-based and (optionally) multi-user feed reader.

If you want to use Tiny Tiny RSS support, don’t forget to activate the external API support in your preferences.

To use Tiny Tiny RSS support, you need to configure a few things. First of all, Newsboat needs to know that you want to use Tiny Tiny RSS and which installation exactly:

urls-source "ttrss"
ttrss-url "https://example.com/ttrss/"

In addition, it requires username and password for authentication:

ttrss-login "myusername"
ttrss-password "mypassword"

Note that double quotes should be escaped, i.e. you should write \" instead of ".

Tiny Tiny RSS provides two modes of usage: single-user mode and multi-user mode. Newsboat needs to know about this, too: in single-user mode, authentication is done via Basic HTTP authentication, while in multi-user mode, authentication is done against Tiny Tiny RSS itself.

ttrss-mode "single"		# "multi" is default

If Tiny Tiny RSS is configured in multi-user mode and still deployed behind an additional HTTP-BasicAuth, the required username and password (which may deviate from ttrss-login and ttrss-password) can be specified in the user-part of the url like this:

ttrss-url "https://htuser:htpasswd@example.com/ttrss/"

With these settings, Newsboat should be able to connect to Tiny Tiny RSS and download your subscribed feeds. Articles or even complete feeds that you marked as read are synchronized directly to Tiny Tiny RSS.

Tiny Tiny RSS provides the ability to "star" and to "publish" articles. Starred articles are basically bookmarks, while published articles can be retrieved via a public RSS feed. Newsboat allows the use of these features by mapping its flags to the "star" and "publish" operations.

In order to use this mapping, you need to configure the flags that shall be used:

ttrss-flag-star "s"
ttrss-flag-publish "p"

After that, use these flags when you edit flags for an article, and these articles will be starred resp. published.

TT-RSS folders are converted into Newsboat tags. You can select and filter feeds by tags; see Tagging and Filter Language for details.

4.6. ownCloud News and nextCloud News

ownCloud News and nextCloud News implement the same protocol, so Newsboat treats them as equivalent. Instructions below apply to both.

First, set your urls-source to ocnews and tell Newsboat where to find your ownCloud instance:

urls-source "ocnews"
ocnews-url "https://localhost/owncloud"

Username and password are required:

ocnews-login "user"
ocnews-password "password"

ownCloud News API uses HTTP basic auth, therefore running ownCloud with SSL is highly recommended. If for any reason you don’t want Newsboat to verify the hostname of your instance against the hostname specified in the SSL certificate you’re using, just say so:

ocnews-verifyhost "no"

If you see intermittent "Authentication failed" errors, try configuring memory caching. That should improve the responsiveness of the API and fix the errors.

ownCloud News provides the ability to "star" articles; starred articles are basically bookmarks. Newsboat allows the use of this feature by mapping user-specified flag to the "star" operation.

In order to use this mapping, you need to configure the flag that shall be used:

ocnews-flag-star "s"

OwnCloud News' folders are converted into Newsboat tags. You can select and filter feeds by tags; see Tagging and Filter Language for details.

4.7. Inoreader

Inoreader is a successor to Google Reader.

To use Newsboat with Inoreader, you should first register your own application. Inoreader’s API doesn’t provide efficient access to all the data Newsboat needs, so the program has to scrape feeds. That generates a lot of API requests, which are rate-limited per application. Thus we ask users to register their own apps, to avoid interfering with others. Please also be mindful of this when enabling auto-reload or lowering reload-time.

With app ID and key in hand, tell Newsboat you’d like to use Inoreader:

urls-source "inoreader"
inoreader-app-id "INSERT_ID_HERE"
inoreader-app-key "INSERT_KEY_HERE"

In addition, Newsboat needs to know your Inoreader username and password so that it can authenticate with Inoreader. Note that this is NOT your login with your Google or Facebook account. If you use one of those to login to Inoreader, you have to create a username and password in Inoreader Preferences > Profile

inoreader-login "your-inoreader-account"
inoreader-password "your-password"

(Note that double quotes should be escaped, i.e. you should write \" instead of ".)

After setting these configuration values, you can start Newsboat. It will authenticate with Inoreader and download your subscription list. If you use "folders" in Inoreader to organize your feeds, Newsboat will make them available via its "tags" capability: each feed is tagged with the name of the folder in which it resides. You can select and filter feeds by tags; see Tagging and Filter Language sections for details.

When you mark single items or complete feeds as read, Newsboat will synchronize this information directly to Inoreader. This, of course, includes opening articles. Toggling read articles back to "unread" is also communicated to Inoreader.

In addition, Inoreader provides the ability to "star" and to "share" articles. Starred articles are basically bookmarks, while shared articles are shown to people that follow your Inoreader account. Newsboat allows the use of this feature by mapping its powerful "flags" to the "star"/"unstar" resp. "share"/"unshare" operations.

In order to use this mapping, all you need to do is to configure the flags that shall be used:

inoreader-flag-share "a"
inoreader-flag-star "b"

After that, use these flags when you edit flags for an article, and these articles will be starred resp. shared.

By default, Newsboat also shows Inoreader "special feeds":

  • Starred items

  • Shared items

  • Liked items

  • Saved web pages

You can disable these feeds by setting the following configuration variable:

inoreader-show-special-feeds no

4.8. Miniflux

Miniflux is a "minimalist and opinionated feed reader" that is self-hostable.

Newsboat can synchronize with Miniflux installations as an alternative to the web front-end.

To configure Miniflux support, set the urls-source option and provide the url to the Miniflux instance to be used:

urls-source "miniflux"
miniflux-url "https://example.com/miniflux/"

In addition, Miniflux requires username and password for authentication:

miniflux-login "myusername"
miniflux-password "mypassword"

Note that double quotes should be escaped, i.e. you should write \" instead of ".

The Miniflux backend will default to using the basic HTTP authentication method, since Newsboat sends extraneous requests when trying to determine the appropriate method automatically. If your Miniflux instance uses a special setup that doesn’t function with basic authentication, it may be necessary to explicitly set the http-auth-method variable in the configuration.

Unlike with other backends, Miniflux’s starring/bookmarking feature is not currently supported in Newsboat.

Miniflux categories are converted into Newsboat tags. You can select and filter feeds by tags; see Tagging and Filter Language for details.

4.9. OPML Online Subscription Mode

This mode provides one-way synchronization of subscriptions from the online service to Newsboat. In other words, Newsboat will know what feeds you’ve subscribed to, but the online service won’t know what feeds you’re reading in Newsboat.

This mode works with any service that publishes your subscriptions in OPML format.

To enable this mode, you need to set an appropriate URLs source and then tell Newsboat where to get the OPML file(s):

urls-source "opml"
opml-url "https://example.com/feeds.opml" "https://example.com/more.opml"

4.10. Passwords for external APIs

There are a number of ways to specify a password, each represented by a separate setting. Newsboat looks for settings in certain order, and uses the first one that it finds. The exact order is described below.

Settings are prefixed by API names; e.g. newsblur-password, feedhq-passwordeval. All APIs support all the settings, so examples below use REMOTEAPI for prefix. Replace it with the name of the remote API you use.

The first setting Newsboat checks is REMOTEAPI-password. It should contain the password in plain text.

The second setting Newsboat checks is REMOTEAPI-passwordfile. It should contain a path to a file; the first line of that file should contain the password in plain text. If the file doesn’t exit, is unreadable, or its first line is empty, Newsboat will exit with an error.

The third setting Newsboat checks is REMOTEAPI-passwordeval. It should contain a command that, when executed, will print out the password to stdout. stderr will be passed through to the terminal.

If the first line of command’s output is empty, or the command fails to execute, Newsboat will exit with an error. This is the most versatile of all the options, because it lets you emulate every other and more; let’s look at it in more detail.

For example, a user might want to store their password in a file encrypted by GPG. They create the file like that:

$ gpg --encrypt --default-recipient-self --output ~/.newsboat/password.gpg

They enter their password, press Enter, and finish the command by pressing ^D (Control and d simultaneously). Then, they specify in their Newsboat config:

REMOTEAPI-passwordeval "gpg --decrypt ~/.newsboat/password.gpg"

Now every time they start Newsboat, GPG will be ran. It’ll probably ask for keyring password, then decrypt the file, and pass its contents to Newsboat, which will use it to authenticate with the remote API.

Note that Newsboat will keep the password in memory the entire time Newsboat is running. Other programs might be able to dump the memory and obtain the password. We don’t currently have any protection from that; patches are welcome.

The user might use any other command here; for example, they could fetch the password from GNOME keyring, KeePass, or somewhere else entirely. The possibilities are truly endless.

If no of the aforementioned settings were found, Newsboat will ask the user for the password using an interactive prompt. If the password that the user enters is empty, Newsboat will give up and exit with an error.

5. Advanced Features

5.1. Tagging

Newsboat comes with the possibility to categorize or "tag", as we call it, RSS feeds. Every RSS feed can be assigned 0 or more tags. Within newsboat, you can then select to only show RSS feeds that match a certain tag. That makes it easy to categorize your feeds in a flexible and powerful way.

Usually, the urls file contains one RSS feed URL per line. To assign a tag to an RSS feed, simply attach it as a single word, separated by blanks such as space or tab. If the tag needs to contain spaces, you must use quotes (") around the tag (see example below). An example urls file may look like this:

https://blog.fefe.de/rss.xml?html interesting conspiracy news "cool stuff"
https://rss.orf.at/news.xml news orf
https://www.heise.de/newsticker/heise.rdf news interesting

When you now start newsboat with this configuration, you can press "t" to select a tag. When you select the tag "news", you will see all three RSS feeds. Pressing "t" again and e.g. selecting the "conspiracy" tag, you will only see the https://blog.fefe.de/rss.xml?html RSS feed. Pressing "^T" clears the current tag, and again shows all RSS feeds, regardless of their assigned tags.

A special type of tag are tags that start with the tilde character (~). When such a tag is found, the feed title is set to the tag name (excluding the ~ character). These type of tags are ignored when any kind of "first tag" property is used. With this feature, you can give feeds any title you want in your feed list:

https://rss.orf.at/news.xml "~ORF News"

Another special type of tag are tags that start with the exclamation mark (!). When such a tag is found, the feed is hidden from the regular list of feeds and its content can only be found through a query feed.

https://rss.orf.at/news.xml "!ORF News (hidden)"

5.2. Scripts and Filters (Snownews Extensions)

Newsboat contains support for Snownews extensions. The RSS feed readers Snownews and Liferea share a common way of extending the readers with custom scripts. Two mechanisms, namely "execurl" and "filter" type scripts, are available and supported by newsboat.

An "execurl" script can be any program that gets executed and whose output is interpreted as RSS feed, while "filter" scripts are fed with the content of a configured URL and whose output is interpreted as RSS feed.

The configuration is simple and straight-forward. Just add to your urls file configuration lines like the following ones:

exec:~/bin/execurl-script
filter:~/bin/filter-script:https://some.test/url

The first line shows how to add an execurl script to your configuration: start the line with exec: and then immediately append the path of the script that shall be executed. If this script requires additional parameters, simply use quotes (see Using Double Quotes for details):

"exec:~/bin/execurl-script param1 param2"

The second line shows how to add a filter script to your configuration: start the line with filter:, then immediately append the path of the script, then append a colon (:), and then append the URL of the file that shall be fed to the script. Again, if the script requires any parameters, simply quote the whole thing:

"filter:~/bin/filter-script param1 param2:https://url/foobar"

In both cases, the tagging feature as described above is still available:

exec:~/bin/execurl-script tag1 tag2 "quoted tag"
filter:~/bin/filter-script:https://some.test/url tag3 tag4 tag5

If you need to write your own extension, see this short guide for an introduction. A collection of existing scripts might also help.

Newsboat comes with an example exec script which shows one way to generate an RSS channel. It also includes a way to see which exact arguments are passed to the script by Newsboat. This example can be found in the doc/examples subdirectory.

5.3. Bookmarking

Newsboat contains a plugin-based bookmarking system. When a user bookmarks a link (possible in the article list, in the article view, and in the URL view), they are asked for the URL to bookmark (already preset with the URL of the current selection), the bookmark title (in most cases preset with the title of the current selection), the bookmark description and (since 2.10) the title of the feed the user is currently in. After the question for the description, an external program, configured via the configuration command bookmark-cmd, is executed with 4 (since 2.10) commandline parameters. The plugin itself implements the actual bookmark saving (e.g. writing the bookmark to an external file, or storing it to a del.icio.us account). If everything went OK, the plugin simply exits. In case something goes wrong while saving the bookmark, it writes out an error message as a single line. This error message is then presented to the user from within newsboat.

Newsboat comes with an example plugin, which implements a simple tab-separated bookmark file. This example can be found in the doc/examples subdirectory.

5.4. Command Line

Like other text-oriented software, Newsboat contains an internal commandline to modify configuration variables ad hoc and to run own commands. It provides a flexible access to the functionality of Newsboat which is especially useful for advanced users.

To start the commandline, type ":". You will see a ":" prompt at the bottom of the screen, similar to tools like vi(m) or mutt. You can now enter commands. Pressing the "Enter" key executes the command (possibly giving feedback to the user) and closes the commandline. You can cancel entering commands by pressing the "Esc" key. The history of all the commands that you enter will be saved to the history.cmdline file, stored next to the cache.db file. The backlog is limited to 100 entries by default, but can be influenced by setting the history-limit configuration variable. To disable history saving, set the history-limit to 0.

The commandline provides you with some help if you can’t remember the full names of commandline commands. By pressing the "Tab" key, newsboat will try to automatically complete your command. If there is more than one possible completion, you can subsequently press the "Tab" key to cycle through all results. If no match is found, no suggestion will be inserted into the commandline. For the set command, the completion also works for configuration variable names.

In addition, some common key combination such as "Ctrl-G" (to cancel input), "Ctrl-K" (to delete text from the cursor position to the end of line), "Ctrl-U" (to clear the whole line) and "Ctrl-W" (to delete the word before the current cursor position) were added.

Please be aware that the input history of both the command line and the search functions are saved to the filesystems, to the files history.cmdline resp. history.search (stored next to the cache.db file). By default, the last 100 entries are saved, but this can be configured (configuration variable history-limit) and also totally disabled (by setting said variable to 0).

Currently, the following command line commands are available:


Syntax: quit
Example: quit

Quit newsboat.


Syntax: q
Example: quit

Alias for quit.


Syntax: save <filename>
Example: save ~/important.txt

Save the currently select article to disk. This works in the article list and in the article view.


Syntax: set <variable>[=<value>|&|!]
Example: set reload-time=15

Set configuration variable <variable> to <value>. If no value is specified, the current value is printed out. Specifying a ! after the name of boolean configuration variables toggles their values, a & directly after the name of a configuration variable of any type resets its value to the documented default value.


Syntax: tag <tagname>
Example: tag news

Only display feeds with the tag <tagname>.


Syntax: goto <case-insensitive substring>
Example: goto foo

Search for a feed whose name contains the case-insensitive substring.


Syntax: source <filename> […​]
Example: source ~/.newsboat/colors

Load the specified configuration files. This allows it to load alternative configuration files or reload already loaded configuration files on-the-fly from the filesystem.


Syntax: dumpconfig <filename>
Example: dumpconfig ~/.newsboat/config.saved

Save current internal state of configuration to file, so that it can be instantly reused as configuration file.


Syntax: dumpform
Example: dumpform

Dump current dialog to text file. This is meant for debugging purposes only.


Syntax: exec <operation>
Example: exec open-all-unread-in-browser-and-mark-read

Run a keybind operation in the current context.


Syntax: number
Example: 30

Jump to the entry with the index <number> (usually seen at the left side of the list). This currently works for the feed list, article list, tag selection and filter selection forms.

5.5. Filter Language

Newsboat provides a powerful filter language that enables the user to filter the content of many dialogs, such as the feed list or the article list. The basic concept is that every feed and every article has a number of attributes which can then be compared with user-supplied values, and these comparisons and be logically AND’ed, OR’ed and grouped.

Examples for simple filter expressions are:

unread_count > 0
rssurl =~ "^https:"
age between 0:10

Logically connecting and grouping such expressions looks like in the following examples:

( unread_count > 0 and unread_count < 10 ) or total_count > 100
( author =~ "Frank" or author =~ "John" ) and ( title =~ "Linux" or title =~ "FreeBSD" )

The possibilities for combining such queries is endless, sky (actually: the available memory) is the limit.

To filter your feeds, press "F" in the feed list, enter your filter expression, and press "Enter". To clear the filter, press "Ctrl-F". To filter the articles in the article list, press "F", enter your expression, and press "Enter". Clearing the filter works the same as before. Be aware that only certain attributes work in both dialogs. The table below lists all available attributes and their context, i.e. an attribute that belongs to a feed can only be matched in the feed list, while an attribute that belongs to an article can only be matched in the article list.

Table 1. Available Comparison Operators
Operator Meaning

= / ==

test for equality

!=

test for inequality; logical negation of the = operator

=~

test whether POSIX extended regular expression matches, case-insensitively

!~

logical negation of the =~ operator

<

less than

>

greater than

<=

less than or equal

>=

greater than or equal

between

within a range of integer values, where the two integer values are separated by a colon (see above for an example)

#

contains; this operator matches if a word is contained in a list of space-separated words (useful for matching tags, see below)

!#

contains not; the negation of the # operator

Table 2. Available Attributes
Attribute Context Meaning

title

article

article title

link

article

article link

author

article

article author

content

article

article body

date

article

publication date of the article

guid

article

a unique identifier of the article

unread

article

indicates whether the article has been read

enclosure_url

article

the URL of an enclosure (e.g. podcast file), empty if there is no enclosure

enclosure_type

article

the MIME type of the enclosure, empty if there is no enclosure

flags

article

The set of flags of the article

age

article

Age of an article (in days)

articleindex

article

Index of an article in an article list

feedtitle

feed, article

title of the feed

description

feed, article

feed description

feedlink

feed, article

link to the feed

feeddate

feed, article

publication date of the feed

rssurl

feed, article

RSS URL of the feed

unread_count

feed, article

number of unread articles in the feed

total_count

feed, article

total number of articles in the feed

tags

feed, article

all tags that are associated with the feed

feedindex

feed, article

Index of a feed in the feed list

Note that it’s also possible to filter for feed attributes when you query for article attributes. This is because every article is internally linked to the feed from which it was downloaded.

5.6. Killfiles

Sometimes, a user is confronted with certain content they don’t want to read, e.g. on topics the user is not interested in or articles written by certain people. In Usenet, such functionality within software is traditionally called a "killfile", i.e. based on the content of this "killfile", articles that match certain conditions do not get displayed and are not presented to the user at all.

In newsboat, such a "killfile" can be implemented on a per-article basis via the configuration file. The most important configuration command for this is ignore-article:

ignore-article "*" "title =~ \"Gentoo\""
ignore-article "https://synflood.at/blog/index.php?/feeds/index.rss2" "title =~ \"newsboat\""

It takes two parameters. The first one is either a URL of a feed, or "*" to match any feed (asterisk is not a pattern, glob or regex—we simply reserve it to mean "all feeds"). The second argument is a filter expression for an article, probably in double quotes to preserve spaces inside. If newsboat hits an article in the specified RSS feed that matches the specified filter expression, then this article is ignored and never presented to the user. The configuration itself can contain as many ignore-article commands as desired.

You can also specify the way an article is ignored. There are two ways available:

  • During download: articles are ignored when a feed is downloaded and parsed, and thus won’t be written to the local cache.

  • During display: articles are downloaded and written to the local cache, but are ignored when a feed is displayed.

Both modes have their advantages and disadvantages: while the download ignore mode saves some storage, you cannot simply "undo" the ignore by removing it from the configuration file: if an ignored article has already vanished from a feed, it won’t reappear. On the other hand, the display ignore mode requires some more space, but has the advantage that an ignore can be "undone" by removing the ignore-article configuration command from the configuration.

The default ignore mode is "download". You can set the ignore-mode in the configuration file:

ignore-mode "display"

5.7. Query Feeds

Query feeds are a mechanism of newsboat to define custom "meta feeds" by using newsboat’s built-in filter language. A query feed is a feed that is aggregated from all currently downloaded articles of all feeds. To narrow down the set of articles, the user has to specify a filter. Only articles that match this filter are added to the query feed.

A query feed is updated whenever it is entered in the feed list. When you change the unread flag of an article, this is reflected in the feed where the article was originally fetched. If you want query feeds to be updated at startup, set prepopulate-query-feeds to yes.

To define a query feed, the user has to add a line to the urls file in the following format:

query:<name of feed>:<filter expression> [<tag> ...]

The query: in the beginning tells newsboat that it’s a query feed, <name of feed> specifies the name under which the query feed shall be displayed in the feed list, and <filter expression> is the filter expression that shall be used. Like every other feed, a query feed can be tagged to organize it like a regular feed.

This feature is often used to create a feed that contains all unread articles:

"query:Unread Articles:unread = \"yes\""

Note the use of double quotes to preserve spaces in the filter expression.

If you want to combine several feeds to one single feed, a good solution is to tag the feeds that you want to combine with one certain tag, and then create a query feed that only displays articles from feeds with that certain tag:

https://domain1.tld/feed.xml fun news tag1
https://domain2.tld/?feed.rss private jokes tag1
https://domain3.tld/feeds.rss news
"query:tag1 Articles:tags # \"tag1\""

In this example, the feeds https://domain1.tld/feed.xml and https://domain2.tld/?feed.rss are aggregated into the query feed named "tag1 Articles", but the feed https://domain3.tld/feeds.rss is not.

Basically, the possibility of what can be realized with query feeds is only limited by what can be queried from articles and feeds with the filter language and by your creativity.

5.8. Flagging Articles

To support custom categorization of articles by the user, it is possible to flag an article. A valid flag is any character from A to Z and from a to z. Every article can be flagged with up to 52 different flags, i.e. every letter from the Roman alphabet in upper and lower case. Flagging is easy: just select an article in the article list, or enter the article view, and press "Ctrl-E". This will start the flag editor. By pressing "Enter", the new flags are saved. You can cancel by pressing the "Esc" key.

The flags of an article can be used in every filter expression. The flags of an article are always ordered, and when new flags are added, ordering is immediately restored. This behaviour can also be relied upon when querying articles via the filter language.

If an article contains one or more flags, it is marked with an "!" in the article list. In the article view, all flags (if available) are listed.

5.9. Macro Support

In newsboat, it’s possible to define macros to execute more than one command at once. A macro is configured using the macro configuration command. The first parameter to macro is the key, all parameters afterwards are operations (as listed in the [available-operations] table above), optionally with parameters on their own, separated by the ; character. Here’s a simple example:

macro k open; reload; quit
macro o open-in-browser; toggle-article-read "read"

When the user presses the macro prefix ("," by default) and then the "k" key, the three operations open, reload and quit will be executed subsequently. The macro prefix can be changed from the default "," to another key, e.g. "+" (if you don’t unbind the default "," you’re left with two macro prefixes):

bind-key + macro-prefix
unbind-key ,

It is also possible to modify configuration variables within macros, which can e.g. be used to temporarily modify the browser configuration variable to do something else, such as running an image viewer from the URLs view:

macro i set browser "feh %u"; open-in-browser; set browser "elinks %u"

You can even use this feature to enqueue any of the URLs from the URLs view to podboat’s download queue (which might be in ~/.locale/share/newsboat/queue if you’re using XDG directories):

macro E set browser "echo %u >> ~/.newsboat/queue"; open-in-browser; set browser "elinks %u"

5.10. Commandline Commands

Newsboat comes with a -x option that indicates that commands added as arguments to the command line shall be executed. Currently, the following commands are available:

  • reload: this option reloads all feeds, and quits newsboat without printing any output. This is useful if a user wants to periodically reload all feeds without always having a running newsboat instance, e.g. from cron.

  • print-unread: this option prints the number of unread articles and quits newsboat. This is useful for users who want to integrate this number into some kind of monitoring system.

5.11. Format Strings

Newsboat contains a powerful format string system to make it possible for the user to configure the format of various aspects of the application, such as the format of entries in the feed list or in the article list.

Format strings are similar to those that are found in the printf function in the C programming language. A format sequence begins with the % character, followed by optional alignment indication: positive numbers indicate that the text that is inserted for the sequence shall be padded on the left to a total width that is specified by the number, while negative number specify padding on the right. Followed by the padding indication comes the actual sequence identifier, which is usually a single letter. %=[width][identifier] centers the sequence, where if w=0 the whole width of the window is used.

In addition, newsboat provides other, more powerful sequences, such as %>[char], which indicates that the text right to the sequence will be aligned right on the screen, and characters between the text on the left and the text on the right will be filled by [char]. Another powerful format is the conditional sequence, %?[char]?[format 1]&[format 2]?: if the text of the sequence identifier [char] contains any non-whitespace characters, then [format 1] will be evaluated and inserted, otherwise [format 2] will be evaluated and inserted. The & and [format 2] are optional, i.e. if the identifier’s text is empty, then an empty string will be inserted.

The following tables show what sequence identifiers are available for which format:

Table 3. Available Identifiers for feedlist-format
Identifier Meaning

d

Feed description

i

Feed index

l

Feed link

L

Feed RSS URL

n

"unread" flag field

S

download status

t

Feed title

T

First tag of a feed in the URLs file

u

"unread/total" field

U

"unread" field

c

"total" field

While a reload-all operation is running, the download status indicates the download status of a feed, which can be "to be downloaded" (indicated by "_"), "currently downloading" (indicated by "."), successfully downloaded (indicated by " ") and "download error" (indicated by "x").

Table 4. Available Identifiers for articlelist-format
Identifier Meaning

a

Article author

D

Publication date. This can be tweaked further with datetime-format

f

Two characters: 1) "N" if article is unread, "D" if article is deleted, a space otherwise; 2) "!" if article has flags, a space otherwise.

n

"unread" field

d

"deleted" field

F

Article flags

i

Article index

t

Article title

T

If the article list displays articles from different feeds, then this identifier contains the title of the feed to which the article belongs.

L

Article length

Table 5. Available Identifiers for selecttag-format
Identifier Meaning

i

Line’s index in the list

T

The tag this line corresponds to

f

Number of unread feeds under this tag

n

Number of unread articles in feeds tagged with this tag

u

Number of feeds tagged with this tag

Table 6. Available Identifiers for notify-format
Identifier Meaning

n

Number of unread articles

f

Number of unread feeds

d

Number of new unread articles (i.e. that were added through the last reload)

D

Number of new unread feeds (i.e. that were added through the last reload)

Table 7. Available Identifiers for podlist-format
Identifier Meaning

i

Download index

d

Currently downloaded size in megabytes, displays one digit of precision

t

Total download size in megabytes, displays one digit of precision

p

Downloaded precentage, displays one digit of precision

k

Download speed, displays two digit of precision, always in KB/s (does not include the "KB/s" text)

K

Download speed, displays two digit of precision, human readable (automatically switches between KB/s, MB/s, and GB/s)

S

Status of download, displays one of the folowing; "queued", "downloading", "ready", "canceled", "deleted", "incomplete", "played", "finished" or "failed"

u

Url of the download

F

Absolute filename of the download from the root directory (e.g. ~/downloads/podcast.mp3 → /home/name/downloads/podcast.mp3)

b

Basename of the download (e.g. /home/name/downloads/podcast.mp3 → podcast.mp3)

Examples:

feedlist-format     "%4i %n %11u %t"
articlelist-format  "%4i %f %D   %?T?|%-17T|  ?%t"
notify-format       "%d new articles (%n unread articles, %f unread feeds)"
podlist-format      "%4i [%-5p %%] %-12S %F"
Table 8. Available Identifiers for download-path and download-filename-format
Identifier Meaning

u

Filename part of the download URL. May be empty. May include a query string

n

Name of the podcast feed

N

Name of the podcast feed. Contains the original feed’s name, even when selected through a query feed

h

Name of the podcast feed’s hostname

t

Title of the podcast episode

e

Extension of the podcast episode

F

Publication date of the podcast episode formatted as yyyy-mm-dd

m

Month when podcast episode was published

b

Abbreviated month name when podcast episode was published

d

Day when podcast episode was published

H

Hour when podcast episode was published

M

Minute when podcast episode was published

S

Second when podcast episode was published

y

Year when podcast episode was published formatted as yy

Y

Year when podcast episode was published formatted as yyyy

5.11.1. Dialog Titles

You can customize the title format of all available dialogs. Here is a list of dialogs with their respective title format configuration variables, and a list of available formats and their meaning. Please note that the title formats are localized, so if you work on a different locale that is supported by newsboat, the actually displayed title text may vary unless you customize it.

Table 9. Dialog Title Formats
Dialog Configuration Variable Default Value

Feed List

feedlist-title-format

%N %V - %?F?Feeds&Your feeds? (%u unread, %t total)%?F? matching filter `%F'&?%?T? - tag `%T'&?

Article List

articlelist-title-format

%N %V - Articles in feed '%T' (%u unread, %t total)%?F? matching filter `%F'&? - %U

Search Result

searchresult-title-format

%N %V - Search results (%u unread, %t total)%?F? matching filter `%F'&?

File Browser

filebrowser-title-format

%N %V - %?O?Open File&Save File? - %f

Directory Browser

dirbrowser-title-format

%N %V - %?O?Open Directory&Save File? - %f

Help

help-title-format

%N %V - Help

Select Tag Dialog

selecttag-title-format

%N %V - Select Tag

Select Filter Dialog

selectfilter-title-format

%N %V - Select Filter

Article View

itemview-title-format

%N %V - Article '%T' (%u unread, %t total)

URL View

urlview-title-format

%N %V - URLs

Dialog List

dialogs-title-format

%N %V - Dialogs
Table 10. Common Title Format Identifiers
Identifier Meaning

N

Name of the program, i.e. "newsboat"

V

Program version

u

Number of unread articles (if applicable)

t

Number of total articles (if applicable)

F

Current filter expression (if applicable, empty if no filter is active)

Table 11. Feed List Title Format Identifiers
Identifier Meaning

T

Currently selected tag (empty if none selected)

Table 12. Article List Title Format Identifiers
Identifier Meaning

T

Feed title

U

Feed URL

Table 13. File Browser Title Format Identifiers
Identifier Meaning

f

Current path

O

Non-empty if file browser is in open mode, empty if in save mode

Table 14. Directory Browser Title Format Identifiers
Identifier Meaning

f

Current path

O

Non-empty if directory browser is in open mode, empty if in save mode

Table 15. Article View Title Format Identifiers
Identifier Meaning

T

Article title

F

Feed title

5.12. Highlighting Text

Newsboat supports the highlighting of text in the feed list, the article list and the article view, using regular expressions to describe patterns to be highlighted. The command syntax goes like this:

highlight <target> <regex> <fgcolor> [<bgcolor> [<attribute> ...]]

Valid values for <target> are feedlist, articlelist, article and all. When specifying all, the matching will be done in all three views. The <regex> must be a POSIX extended regular expression, which will be matched case-insensitive against the text. If multiple highlight matches overlap, the style of the later specified highlight rule will be applied. <fgcolor> and <bgcolor> specify the foreground color resp. the background color of the matches. You can also specify 0 or more attributes. You can find a list of valid colors and attributes in the Configuring Colors.

Examples for possible highlighting configurations are:

highlight all "newsboat" red
highlight article "^(Feed|Title|Author|Link|Date):" default default underline
highlight feedlist "https?://[^ ]+" yellow red bold

Note the use of double quotes to preserve spaces in the regular expressions.

5.12.1. Highlighting Articles in the Article List

In addition to generally highlighting text, there is also a specific way to highlight articles in the article list based on whether they match a certain filter expression. This means that you can highlight items in the article list based on their content. This is done using the highlight-article configuration command.

The syntax is similar to the highlight configuration command, with the difference that there’s no need to specify a target (since it only applies in the article list), and instead of a regular expression, a filter expression is used. After the filter expression, the colors and attributes are specified in the same way.

Example:

highlight-article "author =~ \"Andreas Krennmair\"" white red bold

Note the use of double quotes to preserve spaces in the filter expression.

5.13. Advanced Dialog Management

Newsboat supports an advanced concept of dialogs. Previously, all dialogs (feed list, article list, article view) were internally laid out as a pure stack. In 2.0, this changed: all dialogs are managed in a list, and the user can jump to another, previously opened dialog from everywhere. This allows a user to open more than one article list, more than one article view, etc., and switch between them without closing them.

The main dialog for this feature can be reached by pressing the "v" key. This opens the list of open dialogs. From there, the user can switch to another dialog by selecting the appropriate entry and pressing "Enter", or can close open dialogs by selecting them and pressing "Ctrl-X".

5.14. Podcast Support

A podcast is a media file distributed over the internet using syndication feeds such as RSS, for later playback on portable players or computers. Newsboat contains support for downloading and saving podcasts. This support differs a bit from other podcast aggregators or "podcatchers" in how it is done.

Podcast content is transported in RSS feeds via special tags called "enclosures". Newsboat recognizes these enclosures and stores the relevant information for every podcast item it finds in an RSS feed. Since version 2.0, it also recognizes and handles the Yahoo Media RSS extensions. What the user then can do is to add the podcast download URL to a download queue. Alternatively, newsboat can be configured to automatically do that. This queue is stored in the queue file next to the cache.db file.

The user can then use the download manager podboat to download these files to a directory on the local filesystem. Podboat comes with the newsboat package, and features a look and feel very close to the one of newsboat. It also shares the same configuration file.

Podcasts that have been downloaded but haven’t been played yet remain in the queue but are marked as downloaded. You can remove them by purging them from the queue with the "P" key. After you’ve played a file and close podboat, it will be removed from the queue. The downloaded file remains on the filesystem unless "delete-played-files" is enabled.

6. Podboat Configuration Commands


Syntax: delete-played-files [yes/no]
Default: no
Example: delete-played-files yes

If set to yes, podboat will delete files when their corresponding queue entry is removed (this includes "finished" and "deleted" entries as well).


Syntax: download-path <path>
Default: ~/
Example: download-path "~/Downloads/%h/%n"

Specifies the directory where podboat shall download the files to. Optionally, placeholders can be used to place downloads in a directory structure. See "Format Strings" section of Newsboat manual for details on available formats. This setting is applied at enqueueing time; changing it won’t affect download paths of the podcasts that were already added to the queue.


Syntax: download-filename-format <string>
Default: "%?u?%u&%Y-%b-%d-%H%M%S.unknown?"
Example: download-filename-format "%F-%t.%e"

Specifies how podboat would name the files it downloads (see also download-path). See "Format Strings" section of Newsboat manual for details on available formats.


Syntax: max-downloads <number>
Default: 1
Example: max-downloads 3

Specifies the maximum number of parallel downloads when automatic download is enabled.


Syntax: player <player command>
Default: n/a
Example: player "mp3blaster"

Specifies the player that shall be used for playback of downloaded files.


Syntax: podlist-format <format>
Default: "%4i [%6dMB/%6tMB] [%5p %%] [%12K] %-20S %u → %F"
Example: podlist-format "%i %u %-20S %F"

This variable defines the format of entries in podboat’s download list. See the respective section in the documentation for more information on format strings.

Table 16. Available Operations in Podboat
Operation Default key Description

quit

q

Quit the program.

hard-quit

Q

Quit the program without confirmation.

help

?

Show the help screen.

pb-download

d

Download the currently selected URL.

pb-cancel

c

Cancel the currently selected download.

pb-play

p

Start player with currently selected download.

pb-mark-as-finished

m

Mark currently selected entry as finished.

pb-delete

D

Delete the currently selected URL from the queue.

pb-purge

P

Remove all finished and deleted downloads from the queue and load URLs that were newly added to the queue.

pb-toggle-download-all

a

Toggle the "automatic download" feature where all queued URLs are downloaded one after the other. The "max-downloads" configuration option controls how many downloads are done in parallel.

pb-increase-max-dls

+

Increase the "max-downloads" option by 1.

pb-decrease-max-dls

-

Decrease the "max-downloads" option by 1. If the option is already 1, no further decrease is possible.

A usual "use case" is to configure newsboat to automatically enqueue newly found podcast download URLs. Then, the user reloads the podcast RSS feeds in newsboat, and after that, uses podboat to view the current queue, and either selectively download certain files or automatically download them all together by pressing "a" within podboat.

6.1. Running multiple copies of Newsboat simultaneously

During development and testing, you might want to run a second copy of Newsboat, operating with different config, URLs list or cache file. This can be achieved by creative use of XDG environment variables. This approach is not beautiful, but it works.

First of all, you will need to create a directory to store the data. Let’s call it test:

$ mkdir -p test/newsboat

Note that we also create a subdirectory called newsboat; this is required to satisfy XDG specification. ATTENTION: if this subdirectory is absent, Newsboat will attempt to run on your live data!

You can now create config and urls files inside test/newsboat; you can also copy cache.db if you don’t want to start with a fresh one.

When the files are ready, you can invoke Newsboat as follows:

$ XDG_CONFIG_HOME=test XDG_DATA_HOME=test newsboat

(This will look newsboat binary up in your PATH; if you’ve just built your own, use ./newsboat instead, or provide a full path.)

By modifying the environment in which Newsboat runs you also modify environments of all the programs that Newsboat starts; that includes filters and external HTML renderers. If they rely on XDG_* variables, they will look for things in test directory and might fail.

For filters, you should either copy the files they need to test, or invoke them in such a way that they don’t look in XDG directories at all.

For renderers, you can work around the issue by undoing the modifications to environment, e.g.:

html-renderer "XDG_CONFIG_HOME=$HOME/.config XDG_DATA_HOME=$HOME/.data w3c"

As already said: not beautiful, but gets the job done.

6.2. Using SQLite Triggers with newsboat

This section was written by Elrond, originally for Newsbeuter.

SQLite, the db used by newsboat, supports triggers. These are small snippets of SQL that get executed inside the database by the database engine. They’re stored inside the db and the normal user (including newsboat itself) doesn’t see them. Just the db seems to do some magic: Like changing some values when you change another value.

So what is this good for when looking at newsboat? Well, first off, it’s a hack. The real answer should be to use application logic (do it inside newsboat, not in the db). So: Don’t use this, unless you know, what you’re doing, and unless you have some sort of backup.

6.2.1. Example

So after the "don’t use it" you still want to know, what one can do? So here’s an example.

Suppose you have a strange feed where the articles become "new" by just changing their subject, and nothing else changes. The body is just empty, and the URL keeps the same. This feed really exists. It’s the "updated software rss feed" of some major company and the title just contains the name of the driver and version number. And the URL points to the download page. newsboat considers articles only as new, when they have a new UniqueID (this is good). So those articles are never marked as new (unread) ever again.

So what can we do? We do some magic: We let the db test if newsboat changes the subject and then let itself mark the article again as unread.

  1. You need the sqlite3 command line tool (available via apt-get install sqlite3 on Debian) or some other tool to do direct sql on the sqlite database.

  2. Start sqlite3 with the newsboat db:

    Rivendell:~/.newsboat% sqlite3 cache.db
    SQLite version 3.4.2
    Enter ".help" for instructions
    sqlite>
  3. Create the trigger:

    sqlite> create trigger update_item_title update of title on rss_item
       > for each row when old.title != new.title
       > begin
       >   update rss_item set unread = 1 where rowid == new.rowid;
       > end;
  4. Leave sqlite3 with "Ctrl-D" or .quit.

That’s it. newsboat (well, its db) now marks articles as unread when their title changes. And nicely enough this works all inside newsboat, no need to restart it so that it rereads the cache, that magically modifies itself. It just works.

6.3. Environment variables

BROWSER

Tells Newsboat what browser to use if there is no browser setting in the config file. If this variable doesn’t exist, a default of lynx(1) will be used.

CURL_CA_BUNDLE

Tells Newsboat to use the specified certificate file to verify the peer. The file may contain multiple certificates. The certificate(s) must be in PEM format.

This option is useful if your libcurl is built without useful certificate information, and you can’t rebuild the library yourself.

EDITOR

Tells Newsboat what fallback editor to use when editing the urls file via the edit-urls operation and no VISUAL environment variable is set. If this variable doesn’t exist either, a default of vi(1) will be used.

PAGER

Tells Newsboat what pager to use if the pager setting in the config file is explicitly set to an empty string.

TMPDIR

Tells Newsboat to use the specified directory for storing temporary files. If this variable doesn’t exist, a default of /tmp will be used.

VISUAL

Tells Newsboat what editor to use when editing the urls file via the edit-urls operation. If this variable doesn’t exist, the EDITOR environment variable will be used.

XDG_CONFIG_HOME

Tells Newsboat which base directory to use for the configuration files. See also the section on files for more information.

XDG_DATA_HOME

Tells Newsboat which base directory to use for the data files. See also the section on files for more information.

7. Feedback and security

Please report security vulnerabilities to security@newsboat.org, encrypting your emails to PGP key 4ED6CD61932B9EBE if at all possible.

Non-security issues and general questions can be discussed on the issue tracker and the mailing list.

You can also chat with developers and fellow users on #newsboat at irc.libera.chat (also accessible via webchat and via Matrix) (UPDATED on May 26th 2021 to point to Libera; we do not have a channel on Freenode anymore; UPDATED on May 28th 2021 with links to the webchat and Matrix bridge).

8. License

MIT License

Copyright 2006-2015 Andreas Krennmair <ak@newsbeuter.org>
Copyright 2015-2020 Alexander Batischev <eual.jp@gmail.com>
Copyright 2006-2017 Newsbeuter contributors
Copyright 2017-2020 Newsboat contributors

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.