Simple Sentence Mining (ssmtool) is a universal, cross-platform, multilingual, clipboard-based tool that helps you generate Anki flashcards from whatever source. It is free software under the GNU GPLv3 license, meaning you are free to use, modify, and distribute it, even commerically, as long as it remains under the same license.
It is designed to be simple to set up and straightforward, with the ultimate goal of enabling smooth sentence mining during immersion study. Originally, I wanted to make an Anki add-on, however, I noticed several issues with that idea:
- Anki has no stable API, and addons break regularly on every release, making it difficult to maintain.
- I would like the tool to be extensible for other, potentially better flashcard tools in the future.
Easy word lookups
You can simply copy anything to the clipboard. If it is a single word, it will be looked up right away, which allows you to use it as a dictionary app too.
Otherwise, when the copied content appears in the “Sentence” field, you can simply double click it to look it up. For languages not using spaces, you will have to select it manually, and then either press the button or use the keyboard shortcut. Better support for Chinese and Japanese is expected in the near future.
Online and local dictionaries
From version 0.2.0 ssmtool comes with three dictionaries out of the box
- English Wiktionary
wikt-en(Definitions are in English; support for other languages here depends on progress on their API)
- Google Translate
gtrans(Translating to any other supported language)
- Google Dictionary
gdict(Monolingual, i.e., definitions in the same language)1
Local dictionaries (StarDicts, Migaku JSON files, and simple JSON key:value pairs) are also supported. To find dictionaries, check this post.
Frequency lists and pronunciations
SSM is able to import frequency lists in the format of a JSON array file, such as the ones available on the MEGA folder linked in the post linked above. If a frequency list is enabled, frequency rank for words will be displayed on the main window to help you decide whether to add a note.
In addition, SSM can fetch native pronunciations from Forvo and export the audio to Anki.
Many similar tools such as LingQ, Readlang, and Learn with Texts do not have the ability to lemmatize words, which is to convert words into their dictionary forms. The issue with this is that many dictionaries do not have all forms of a word listed, especially in highly inflectional languages (Russian, Finnish). With lemmatization, dictionary lookups become much smoother because there is no longer a need to manually modify the word before looking it up.
This program leverages AnkiConnect to add cards directly to Anki, without using csv exports. Tags and custom note layouts are also supported.
This extension enables you to mine sentences with a single click on most websites, further smoothening the immersion experience.
Language Support Matrix
There are three components you need to install to make this program work.
Main Desktop Application
Arch Linux AUR:
Otherwise, you can use
pip install ssmtool (or pip3 if appropriate)
This will install a desktop file which you should be able to see from your
launcher menu. If you do not use a desktop environment, you can launch it
through the command line
On Linux the appearance of the app depends on the system Qt theme.
Windows and macOS
Even though we encourage the use of free software, nonfree operating systems are still supported by this tool. You can go to the Github Releases for standalone versions.
If you are feeling brave, you are also welcome to go to AppVeyor to obtain the latest builds, but they are not guaranteed to run.
Alternatively, or if you would like a development version, you can also run
pip command above.
Only 64 bit Windows 7+ is supported. On Windows you need the Microsoft Visual C++ Redistributable Package, which you may or may not already have.
Since I do not have a Mac computer now, the macOS builds are not tested. If you use a Mac and found that it does not work, please report to Github or our chatroom. If you are interested you can also help test future versions.
Anki Add-on (Required for card creation)
Download and install Anki dekstop (Not mobile or Anki Universal). Skip if you already installed it.
Then, install the AnkiConnect addon. You do not have to change any settings for it.
Mac users: You must have Anki open on the foreground (i.e. visible on your desktop), or otherwise disable the App Nap feature. If you do not do this, AnkiConnect will not respond and will cause this program to be very slow and/or unresponsive.
Browser (Optional but highly recommended)
Install extension for your browser:
Alternatively, if you have articles or books to read, you can also use the HTML reader app, which does not require using the browser extension.
You need to first select a target language from the list. Then, you can select a dictionary. We recommend using Google translation only if the other two are not available, because translations are always less detailed than dictionary definitions and may not provide the full range of meanings needed.
We recommend leaving lemmatization on as it is by default. It can greatly boost dictionary coverage for many languages.
Optionally, you can add frequency lists and local dictionary files via the bottom option of “Manage local dictionaries”. Consult this post for compatible files.
Next, on the Anki tab, you will see a number of settings. You usually do not have to change the first one, which is the API endpoint, unless you configured a different endpoint in AnkiConnect, but in that case you will know how to do this. You should then select a deck, note type, and match data fields into note fields. To do this you must have a note type with at least three fields, one each for Sentence, Word, and Definition.
If you do not have one, you can download it from here and import it into Anki. Then, select “ssmtool-reading” and simply match the field names.
You’re done! Now you are ready to mine sentences.
When you see any sentence from anywhere, you can simply copy it to clipboard. It will appear on the Sentence field right away. Then, double click on any word, and you will get the definition. You can look up words from the Definition field too. Then, when you are satisfied with the data, click on Add Note to send it to Anki.
Note: It seems that on macOS clipboard change may not always be detected. In that case simply use the “Read clipboard” button.
When you turn on the extension, you will notice that sentences are underlined in green (will become configurable in the future). Whenever you click on any word, ssmtool will receive both the whole sentence and the word under your cursor. The word will be looked up immediately too. Chances are, with lemmatization on, this is exactly the word you want. In that case, just press Ctrl/Cmd + S, and you can keep reading!
You can use this tool in combination with mpv with the mpvacious plugin to make it copy subtitles to clipboard continuously. Audio data is currently not supported, but may be added in the near future either as an extension to ssmtool, or as a standalone mpv plugin.
Currently ssmtool does not support directly adding audio during card creation.
However you can easily get the audio during review with
AwesomeTTS Anki addon.
it, you can go to
Tools > Manage Note Types > (select your note type) > Cards > Add TTS
Then, select the Sentence field, the Word field, or both. Refold recommends
putting audio on the back of the card, but it is also possible to put it on
the front side.
You are welcome to report bugs, suggest features/enhancements, or ask for clarifications by opening a GitHub issue.
You are also welcome ask _z#6358 on Refold Discord for help.
The maintainer of the project has said that they would transition to English Wiktionary definitions due to copyright concerns. So far, this has not happened. In the event that it does, this option will be removed, as there is already an option for using English Wiktionary. ↩︎
“Full support” simply means the full feature-set is available. Usability can depend on a lot of other factors, such as dictionary coverage and lemmatizer accuracy, both of which can vary depending on the language. ↩︎
Only the Brazillian variant. It will be selected automatically if Google dictionary is chosen ↩︎