Skip to content
Switch branches/tags


Failed to load latest commit information.
Latest commit message
Commit time


A command line tool for remembering what you were doing and tracking what you've done.

If you're one of the rare people like me who find this useful, feel free to buy me some coffee.


What and why

doing is a basic CLI for adding and listing "what was I doing" reminders in a TaskPaper-formatted text file. It allows for multiple sections/categories and flexible output formatting.

While I'm working, I have hourly reminders to record what I'm working on, and I try to remember to punch in quick notes if I'm unexpectedly called away from a project. I can do this just by typing doing now tracking down the CG bug.

If there's something I want to look at later but doesn't need to be added to a task list or tracker, I can type doing later check out the pinboard bookmarks from macdrifter. When I get back to my computer --- or just need a refresher after a distraction --- I can type doing last to see what the last thing on my plate was. I can also type doing recent (or just doing) to get a list of the last few entries. doing today gives me everything since midnight for the current day, making it easy to see what I've accomplished over a sleepless night.

Side note: I actually use the library behind this utility as part of another script that mirrors entries in Day One that have the tag wwid. I can use the hourly writing reminders and enter my stuff in the quick entry popup. Someday I'll get around to cleaning that up and putting it out there.


$ [sudo] gem install doing

To install the latest version, use --pre:

$ [sudo] gem install --pre doing

Only use sudo if your environment requires it. If you're using the system Ruby on a Mac, for example, it will likely be necessary. If gem install doing fails, then run sudo gem install doing and provide your administrator password.

Run doing config to open your ~/.doingrc file in the editor defined in the $EDITOR environment variable. Set up your doing_file right away (where you want entries to be stored), and cover the rest after you've read the docs.

See the support section below for troubleshooting details.

The "doing" file

The file that stores all of your entries is generated the first time you add an entry (with doing now or doing later). By default, the file is created in ~/, but you can modify this in the config file.

The format of the "doing" file is TaskPaper-compatible. You can edit it by hand at any time (in TaskPaper or any text editor), but it uses a specific format for parsing, so be sure to maintain the dates and pipe characters.

Notes are anything in the list without a leading hyphen and date. They belong to the entry directly before them, and they should be indented one level beyond the parent item.

When using the now and later commands on the command line, you can start the entry with a quote and hit return, then type the note and close the quote. Anything after the first line will be turned into a TaskPaper-compatible note for the task and can be displayed in templates using %note.

Notes can be prevented from ever appearing in output with the global option --no-notes: doing --no-notes show all.

Auto tagging (adding tags listed in .doingrc under autotag and default_tags) can be skipped for an entry with the -x global option: doing -x done skipping some automatic tagging.


A basic configuration looks like this:

doing_file: /Users/username/Dropbox/doing.taskpaper
current_section: Currently
default_template: '%date: %title%note'
default_date_format: '%Y-%m-%d %H:%M'
marker_tag: flagged
marker_color: yellow
default_tags: []
editor_app: TextEdit
:include_notes: true
    date_format: '%F %_I:%M%P'
    section: Currently
    count: 10
    wrap_width: 0
    template: '%boldblack%date %boldgreen| %boldwhite%title%default%note'
    order: desc
    date_format: '%Y-%m-%d %H:%M'
    template: '%date | %title%note'
    wrap_width: 0
    date_format: '%_I:%M%P'
    template: '%date: %title%odnote'
    wrap_width: 0
    date_format: '%_I:%M%P on %a'
    template: '%title (at %date)%odnote'
    wrap_width: 0
    date_format: '%_I:%M%P'
    template: '%date > %title%odnote'
    wrap_width: 50
  - coding
  - design
    - thinking
    - idea

The config file is stored in ~/.doingrc, and a skeleton file is created on the first run. Just run doing on its own to create the file.

Per-folder configuration

Any options found in a .doingrc anywhere in the hierarchy between your current folder and your home folder will be appended to the base configuration, overriding or extending existing options. This allows you to put a .doingrc file into the base of a project and add specific configurations (such as default tags) when working in that project on the command line. These can be cascaded, with the closest .doingrc to your current directory taking precedence, though I'm not sure why you'd want to deal with that.

Possible uses:

  • Define custom HTML output on a per-project basis using the html_template option for custom templates. Customize time tracking reports based on project or client.
  • Define default_tags for a project so that every time you doing now from within that project directory or its subfolders, it gets tagged with that project automatically.

Any part of the configuration can be copied into these local files and modified. You only need to include the parts you want to change or add.

Doing file location

The one thing you'll probably want to adjust is the file that the notes are stored in. That's the doing_file key:

doing_file: /Users/username/Dropbox/nvALT2.2/?? What was I

I keep mine in my nvALT folder for quick access and syncing between machines. If desired, you can give it a .taskpaper extension to make it more recognizable to other applications. (If you do that in nvALT, make sure to add taskpaper as a recognized extension in preferences).

"Current actions" section

You can rename the section that holds your current tasks. By default, this is Currently, but if you have some other bright idea, feel free:

current_section: Currently

Default editors

The setting editor_app only applies to Mac OS X users. It's the default application that the command doing open will open your WWID file in. If this is blank, it will be opened by whatever the system default is, or you can use -a app_name or -b bundle_id to override.

In the case of the doing now -e command, your $EDITOR environment variable will be used to complete the entry text and notes. Set it in your ~/.bash_profile or whatever is appropriate for your system:

export EDITOR="mate -w"

The only requirements are that your editor be launchable from the command line and able to "wait." In the case of Sublime Text and TextMate, just use -w like this: export EDITOR="subl -w".


The config also contains templates for various command outputs. Include placeholders by placing a % before the keyword. The available tokens are:

  • %title: the "what was I doing" entry line
  • %date: the date based on the template's date_format setting
  • %shortdate: a custom date formatter that removes the day/month/year from the entry if they match the current day/month/year
  • %note: Any note in the entry will be included here, a newline and tabs are automatically added.
  • %odnote: The notes with a leading tab removed (outdented note)
  • %chompnote: Notes on one line, beginning and trailing whitespace removed.
  • %section: The section/project the entry is currently in
  • %hr: a horizontal rule (-) the width of the terminal
  • %hr_under: a horizontal rule (_) the width of the terminal
  • %n: inserts a newline
  • %t: inserts a tab
  • %[color]: color can be black, red, green, blue, yellow, magenta, cyan, or white
    • you can prefix bg to affect background colors (%bgyellow)
    • prefix bold and boldbg for strong colors (%boldgreen, %boldbgblue)
    • there are some random special combo colors. Use doing colors to see the list
  • %interval: when used with the -t switch on the show command, it will display the time between a timestamp or @start(date) tag and the @done(date) tag, if it exists. Otherwise, it will remain empty.

Date formats are based on Ruby strftime formatting.

My normal template for the recent command looks like this:

  date_format: '%_I:%M%P'
  template: '%date > %title%odnote'
  wrap_width: 88

And it outputs:

$ doing recent 3
 4:30am > Made my `console` script smarter...
    Checks first argument to see if it's a file, if it is, that's the log

    Else, it checks the first argument for a ".log" suffix and does a search in the user
    application logs with `find` for it.

    Otherwise, system.log.

    I also made an awesome Cope wrapper for it...
12:00pm > Working on `doing` again.
12:45pm > I think this thing (doing) is ready to document and distribute

You can get pretty clever and include line breaks and other formatting inside of double quotes. If you want multiline templates, just use \n in the template line, and after the next run it will be rewritten as proper YAML automatically.

For example, this block:

  date_format: '%_I:%M%P'
  template: "\n%hr\n%date\n > %title%odnote\n%hr_under"
  wrap_width: 100

will rewrite to:

  date_format: '%_I:%M%P'
  template: |2-

     > %title%odnote
  wrap_width: 100

and output my recent entries like this:

$ doing recent 3
 > Made my `console` script smarter...
    Checks first argument to see if it's a file, if it is, that's the log

    Else, it checks the first argument for a ".log" suffix and does a search in the user application
    logs with `find` for it.

    Otherwise, system.log.

    I also made an awesome Cope wrapper for it...

 > Working on `doing` again.

 > I think this thing (doing) is ready to document and distribute


Custom views

You can create your own "views" in the ~/.doingrc file and view them with doing view view_name. Just add a section like this:

    section: Old
    count: 5
    wrap_width: 0
    date_format: '%F %_I:%M%P'
    template: '%date | %title%note'
    order: asc
    tags: done finished cancelled
    tags_bool: ANY

You can add additional custom views. Just nest them under the views key (indented two spaces from the edge). Multiple views would look like this:

    section: Later
    count: 5
    wrap_width: 60
    date_format: '%F %_I:%M%P'
    template: '%date | %title%note'
    section: Old
    count: 5
    wrap_width: 0
    date_format: '%F %_I:%M%P'
    template: '%date | %title%note' 

The section key is the default section to pull entries from. Count and section can be overridden at runtime with the -c and -s flags. Setting section to All will combine all sections in the output.

You can add new sections with doing add_section section_name. You can also create them on the fly by using the -s section_name flag when running doing now. For example, doing now -s Misc just a random side note would create the "just a random side note" entry in a new section called "Misc," if Misc didn't already exist.

The tags and tags_bool keys allow you to specify tags that the view is filtered by. You can list multiple tags separated by spaces, and then use tags_bool to specify ALL, ANY, or NONE to determine how it handles the multiple tags.

The order key defines the sort order of the output. This is applied after the tasks are retrieved and cut off at the maximum number specified in count.

Regarding colors, you can use them to create very nice displays if you're outputting to a color terminal. Example:

  date_format: '%F %_I:%M%P'
  section: Currently
  count: 10
  wrap_width: 0
  template: '%boldblack%date %boldgreen| %boldwhite%title%default%note'


You can also specify a default output format for a view. Most of the optional output formats override the template specification (html, csv, json). If the view command is used with the -o flag, it will override what's specified in the file.


You can use the following colors in view templates. Set a foreground color with a named color:


You can also add a background color (%bg[color]) by placing one after the foreground color:


There are bold variants for both foreground and background colors

... etc.

... etc.

And a few special colors you'll just have to try out to see (or just run doing colors):


Any time you use one of the foreground colors it will reset the bold and background settings to their default automatically. You can force a reset to default terminal colors using %default.

HTML Templates

For commands that provide an HTML output option, you can customize the templates used for markup and CSS. The markup uses HAML, and the styles are pure CSS.

To export the default configurations for customization, use doing templates --type=[HAML|CSS]. This will output to STDOUT where you can pipe it to a file, e.g. doing templates --type=HAML > my_template.haml. You can modify the markup, the CSS, or both.

Once you have either or both of the template files, edit .doingrc and look for the html_template: section. There are two subvalues, haml: and css:. Add the path to the templates you want to use. A tilde may be substituted for your home directory, e.g. css: ~/styles/doing.css.


Keywords in your entries can trigger automatic tagging, just to make life easier. There are three tools available: default tags, whitelisting, and synonym tagging.

Default tags are tags that are applied to every entry. You probably don't want to add these in the root configuration, but using a local .doingrc in a project directory that defines default tags for that project allows anything added from that directory to be tagged automatically. A local .doingrc in my Marked development directory might contain:

default_tags: [marked,coding]

And anything I enter while in the directory gets tagged with @marked and @coding.

A whitelist is a list of words that should be converted directly into @tags. If my whitelist contains "design" and I type doing now working on site design, that's automatically converted to "working on site @design."

Synonyms allow you to define keywords that will trigger their parent tag. If I have a tag called @design, I can add "typography" as a synonym. Then entering doing now working on site typography will become "working on site typography @design."

White lists and synonyms are defined like this:

    - typography
    - layout
    - thinking
    - idea
  - brainstorming
  - coding

Note that you can include a tag with synonyms in the whitelist as well to tag it directly when used.


doing [global options] command [command options] [arguments...]

Global options:

-f, --doing_file=arg - Specify a different doing_file (default: none)
--help               - Show this message
--[no-]notes         - Output notes if included in the template (default: enabled)
--stdout             - Send results report to STDOUT instead of STDERR
--version            - Display the program version
-x, --[no-]noauto    - Exclude auto tags and default tags


help           - Shows a list of commands and global options
help [command] - Shows help for any command (`doing help now`)

Adding entries:

now, did - Add an entry
later    - Add an item to the Later section
done     - Add a completed item with @done(date). No argument finishes last entry.
meanwhile - Finish any @meanwhile tasks and optionally create a new one

The doing now command can accept -s section_name to send the new entry straight to a non-default section. It also accepts --back=AMOUNT to let you specify a start date in the past using "natural language." For example, doing now --back=25m ENTRY or doing now --back="yesterday 3:30pm" ENTRY.

If you want to use --back with doing done but want the end time to be different than the start time, you can either use --took in addition, or just use --took on its own as it will backdate the start time such that the end time is now and the duration is equal to the value of the --took argument.

You can finish the last unfinished task when starting a new one using doing now with the -f switch. It will look for the last task not marked @done and add the @done tag with the start time of the new task (either the current time or what you specified with --back).

doing done is used to add an entry that you've already completed. Like now, you can specify a section with -s section_name. You can also skip straight to Archive with -a.

doing done can also backdate entries using natural language with --back 15m or --back "3/15 3pm". That will modify the starting timestamp of the entry. You can also use --took 1h20m or --took 1:20 to set the finish date based on a "natural language" time interval. If --took is used without --back, then the start date is adjusted (--took interval is subtracted) so that the completion date is the current time.

When used with doing done, --back and --took allow time intervals to be accurately counted when entering items after the fact. --took is also available for the doing finish command, but cannot be used in conjunction with --back. (In finish they both set the end date, and neither has priority. --back allows specific days/times, --took uses time intervals.)

All of these commands accept a -e argument. This opens your command line editor (as defined in the environment variable $EDITOR). Add your entry, save the temp file, and close it. The new entry is added. Anything after the first line is included as a note on the entry.

doing meanwhile is a special command for creating and finishing tasks that may have other entries come before they're complete. When you create an entry with doing meanwhile [entry text], it will automatically complete the last @meanwhile item (dated @done tag) and add the @meanwhile tag to the new item. This allows time tracking on a more general basis, and still lets you keep track of the smaller things you do while working on an overarching project. The meanwhile command accepts --back [time] and will backdate the @done tag and start date of the new task at the same time. Running meanwhile with no arguments will simply complete the last @meanwhile task.

See doing help meanwhile for more options.

Modifying entries:

finish      - Mark last X entries as @done
tag         - Tag last entry
note        - Add a note to the last entry

doing finish by itself is the same as doing done by itself. It adds @done(timestamp) to the last entry. It also accepts a numeric argument to complete X number of tasks back in history. Add -a to also archive the affected entries.

doing finish also provides an --auto flag, which you can use to set the end time of any entry to 1 minute before the start time of the next. Running a command such as doing finish --auto 10 will go through the last 10 entries and sequentially update any without a @done tag with one set to the time just before the next entry in the list.

As mentioned above, finish also accepts --back "2 hours" (sets the finish date from time now minus interval) or --took 30m (sets the finish date to time started plus interval) so you can accurately add times to completed tasks, even if you don't do it in the moment.

Tagging and Autotagging

tag adds one or more tags to the last entry, or specify a count with -c X. Tags are specified as basic arguments, separated by spaces. For example:

doing tag -c 3 client cancelled

... will mark the last three entries as @client @cancelled. Add -r as a switch to remove the listed tags instead.

You can optionally define keywords for common tasks and projects in your .doingrc file. When these keywords appear in an item title, they'll automatically be converted into @tags. The whitelist tags are exact (but case insensitive) matches.

You can also define synonyms, which will add a tag at the end based on keywords associated with it. When defining synonym keys, be sure to indent but not hyphenate the keys themselves, while hyphenating the list of synonyms at the same indent level as their key. See playing and writing in the example below for illustration. Follow standard YAML syntax.

To add autotagging, include a section like this in your ~/.doingrc file:

  - doing
  - mindmeister
  - marked
  - playing
  - working
  - writing
    - hacking
    - tweaking
    - toying
    - messing
    - blogging
    - posting
    - publishing
Tag transformation

You can include a transform section in the autotag config which contains pairs of regular expressions and replacement patterns separated by a colon. These will be used to look at existing tags in the text and generate additional tags from them. For example:

autotag: transform: - (\w+)-\d+:$1

This creates a search pattern looking for a string of word characters followed by a hyphen and one or more digits, e.g. @projecttag-12. Do not include the @ symbol in the pattern. The replacement ($1) indicates that the first matched group (in parenthesis) should be used to generate the new tag, resulting in @projecttag being added to the entry.


note lets you append a note to the last entry. You can specify a section to grab the last entry from with -s section_name. -e will open your $EDITOR for typing the note, but you can also just include it on the command line after any flags. You can also pipe a note in on STDIN (echo "fun stuff"|doing note). If you don't use the -r switch, new notes will be appended to the existing notes, and using the -e switch will let you edit and add to an existing note. The -r switch will remove/replace a note; if there's new note text passed when using the -r switch, it will replace any existing note. If the -r switch is used alone, any existing note will be removed.

You can also add notes at the time of entry by using the -n or --note flag with doing now, doing later, or doing done. If you pass text to any of the creation commands which has multiple lines, everything after the first line break will become the note.

Displaying entries:

show      - List all entries
recent    - List recent entries
today     - List entries from today
yesterday - List entries from yesterday
last      - Show the last entry
grep      - Show entries matching text or pattern

doing show on its own will list all entries in the "Currently" section. Add a section name as an argument to display that section instead. Use "all" to display all entries from all sections.

You can filter the show command by tags. Simply list them after the section name (or all). The boolean defaults to ANY, meaning any entry that contains any of the listed tags will be shown. You can use -b ALL or -b NONE to change the filtering behavior: doing show all done cancelled -b NONE will show all tasks from all sections that do not have either @done or @cancelled tags.

Use -c X to limit the displayed results. Combine it with -a newest or -a oldest to choose which chronological end it trims from. You can also set the sort order of the output with -s asc or -s desc.

The show command can also show the time spent on a task if it has a @done(date) tag with the -t option. This requires that you include a %interval token in template -> default in the config. You can also include @start(date) tags, which override the timestamp when calculating the intervals.

If you have a use for it, you can use -o csv on the show or view commands to output the results as a comma-separated CSV to STDOUT. Redirect to a file to save it: doing show all done -o csv > ~/Desktop/done.csv. You can do the same with -o json.

doing yesterday is great for stand-ups (thanks to Sean Collins for that!). Note that you can show yesterday's activity from an alternate section by using the section name as an argument (e.g. doing yesterday archive).

doing on allows for full date ranges and filtering. doing on saturday, or doing on one month to today will give you ranges. You can use the same terms with the show command by adding the -f or --from flag. doing show @done --from "monday to friday" will give you all of your completed items for the last week (assuming it's the weekend).

You can also show entries matching a search string with doing grep (synonym doing search). If you want to search with regular expressions or for an exact match, surround your search query with forward slashes, e.g. doing search /project name/. If you pass a search string without slashes, it's treated as a fuzzy search string, meaning matches can be found as long as the characters in the search string are in order and with no more than three other characters between each. By default searches are across all sections, but you can limit it to one with the -s SECTION_NAME flag. Searches can be displayed with the default template, or output as HTML, CSV, or JSON.


view     - Display a user-created view
views    - List available custom views

Display any of the custom views you make in ~/.doingrc with the view command. Use doing views to get a list of available views. Any time a section or view is specified on the command line, fuzzy matching will be used to find the closest match. Thus, lat will match Later, etc..


sections    - List sections
choose      - Select a section to display from a menu
add_section - Add a new section to the "doing" file


archive  - Move entries between sections
open     - Open the "doing" file in an editor (OS X)
config   - Edit the default configuration


    -k, --keep=arg - Count to keep (ignored if archiving by tag) (default: 5)
    -t, --to=arg   - Move entries to (default: Archive)
    -b, --bool=arg - Tag boolean (default: AND)

The archive command will move entries from one section (default: Currently) to another section (default: Archive).

doing archive on its own will move all but the most recent 5 entries from currently into the archive.

doing archive other_section will archive from other_section to Archive.

doing archive other_section -t alternate will move from other_section to alternate. You can use the -k flag on any of these to change the number of items to leave behind. To move everything, use -k 0.

You can also use tags to archive. You define the section first, and anything following it is treated as tags. If your first argument starts with @, it will assume all sections and assume any following arguments are tags.

By default, tag archiving uses an AND boolean, meaning all the tags listed must exist on the entry for it to be moved. You can change this behavior with -b OR or -b NONE (ALL and ANY also work).

Example: Archive all Currently items for @client that are marked @done

doing archive @client @done


Shell completion

Bash: See the file doing.completion.bash in the git repository for full bash completion. Thanks to fcrespo82 for getting it started.

Zsh: See the file doing.completion.zsh in the git repository for zsh completion. Courtesy of Gabe Anzelini.

Fish: See the file in the git repository for Fish completion. This is the least complete of all of the completions, but it will autocomplete the first level of subcommands, and your custom sections and views for the doing show and doing view commands.


The LaunchBar action requires that doing be available in /usr/local/bin/doing. If it's not (because you're using RVM or similar), you'll need to symlink it there. Running the action with Return will show the latest 9 items from Currently, along with any time intervals recorded, and includes a submenu of Timers for each tag.

Pressing Spacebar and typing allows you to add a new entry to currently. You an also trigger a custom show command by typing "show [section/tag]" and hitting return.

Point of interest, the LaunchBar Action makes use of the -o json flag for outputting JSON to the action's script for parsing.

See for the download.

Evan Lovely has created an Alfred workflow as well.


Errors after "Successfully installed..."

If you get errors in the terminal immediately after a message like:

Successfully installed doing-x.x.x
2 gems installed may just be documentation related. If running doing works, you can ignore them. If not, try running the install command again with --no-document:

$ gem install --no-document doing

Command not found

If running doing after a successful install gives you a "command not found" error, then your gem path isn't in your $PATH, meaning the system can't find it. To locate the gem and link it into your path, you can try this:

cd $GEM_PATH/bin
ln -s doing /usr/local/bin/

Then try running doing and see if it works.

Encoding errors

Ruby is rife with encoding inconsistencies across platforms and versions. Feel free to file issues (see below).


As a free project, doing isn't heavily supported, but you can get support from myself and other users on GitHub. If you run into a replicatable bug in your environment, please post an issue and include your platform, OS version, and the result of ruby -v, along with a copy/paste of the error message. To get a more verbose error message, try running GLI_DEBUG=true doing [...] for a full trace.

Please try not to email me directly about GitHub projects.

Developer notes

Feel free to poke around, I'll try to add more comments in the future (and retroactively).

{% donate doing now sending coffee money to Brett. %}