Quick Answer: Word Atlas (词光) is a free, open-source IELTS vocabulary app with 7,298 illustrated word entries, UK and US audio, and FSRS-based spaced review. It needs no account, keeps progress on your device, and its dataset is licensed CC BY 4.0. It's good for building recognition, but it isn't an official IELTS word list, and its content hasn't been manually proofread entry by entry yet.
Lexical Resource is one of four equally weighted criteria in the IELTS Writing and Speaking band descriptors, so vocabulary gaps cost you points on two of the four test components. Most candidates respond by buying a thick word list and grinding through it. That's where things stall. Word Atlas takes a different path: it pairs each word with an illustration, native-speaker audio, and a review schedule. Here's what it does well, where it falls short, and how to use it without wasting thousands of entries.
What Word Atlas actually is
Word Atlas is a bilingual (Chinese and English) study app built around the idea that a picture helps you hold a word's meaning. The source project lists 7,298 IELTS-oriented entries spread across 19 topics and CEFR levels A1 to C2. Each entry includes English and Chinese senses, two original example sentences with translations, collocations, word-family notes, and a usage tip. The README is in Chinese, with an English version alongside it.
The codebase is open source under MIT, and the dataset is released under CC BY 4.0, so you can reuse the words in your own tools. You can browse the code and data on GitHub, or try the hosted version at 0xlei.tech. The two don't match exactly. The project says the online site iterates independently, so the features you see there may differ from the repo.
How the picture-first learning loop works
The app runs a four-step loop on each word. Here's the sequence as the project describes it:
- See: Read the illustration, meaning, and example sentences together.
- Hear: Play UK or US pronunciation and a recorded example sentence.
- Recall: Answer without seeing the word, through meaning matching, picture-based sentence completion, listening discrimination, or spelling.
- Review: Let the scheduler decide when you'll see the word again.
The "Recall" step matters most. Passive exposure feels productive, but the testing effect, one of the most replicated findings in learning research, says that pulling information out of memory strengthens it more than re-reading does. Word Atlas builds retrieval into the default flow, which is the part most flashcard-free word lists skip.
Dual coding, the idea that pairing words with images creates two retrieval routes, is also well studied. The illustrations are a reasonable bet. Here's the catch: a picture only helps if you still have to produce the word from it. If you just look at the image and nod, you've bought a decoration, not a memory aid.
Why the scheduler matters more than the artwork
Here's the part most app reviews skip. The illustrations get all the attention, but the scheduling engine does most of the long-term work. Word Atlas uses FSRS (Free Spaced Repetition Scheduler), via the TypeScript library ts-fsrs. FSRS models each card's stability and difficulty, then estimates when your recall probability will drop to a target level. The project lets you adjust that target.
Why does this beat fixed intervals? Older systems such as the SM-2 algorithm used by early Anki versions apply fixed multipliers to every card. FSRS fits its parameters to your own review history, so a word you forget every time gets shown sooner, and a word you always recall gets pushed out further. Anki added FSRS as an option in version 23.10, so you can compare the two approaches side by side if you already use Anki.
The counter-intuitive takeaway: a plain flashcard app with a good scheduler will often beat a beautiful app with a weak one. Word Atlas gets both, which is why the scheduling claims are worth reading in the project's learning-model documentation before you commit.
Where the open dataset earns its keep
The open data is the strongest case for this project, even if you never open the app. The repository includes:
- Word entries: 7,298 records in JSON and CSV, tagged with CEFR level and topic.
- Example sentences: 14,596 sentences with Chinese translations.
- Illustrations: 7,298 512 px WebP images, plus thumbnails.
- Audio: 22,905 clips covering British and American pronunciation and sentence readings.
If you teach or build study tools, you can filter the data directly. For example, you could pull every B2 word tagged with the environment topic into a custom deck. Here's the practical use case:
- Build an Anki deck from the CSV, using your own image set.
- Generate a printed vocabulary sheet for a classroom, with CEFR levels preserved.
- Audit your existing word list against the CEFR tags to find gaps.
The CC BY 4.0 license requires attribution to "Word Atlas contributors (词光)", so keep that credit visible in anything you publish.
Comparison: Word Atlas vs Anki vs a paper word list
| Feature | Word Atlas | Anki (desktop) | Paper word list |
|---|---|---|---|
| Cost and licence | Free; MIT code, CC BY 4.0 data | Free; open-source desktop app | Free or purchased once |
| Account needed | No | No (sync is optional) | No |
| Illustrations | Yes, 7,298 original images | Only if you add them | No |
| Audio | UK and US for every entry, plus sentence audio | Only if you add it | No |
| Scheduler | FSRS, with an adjustable retention target | FSRS (since 23.10) or SM-2 | None; you schedule yourself |
The table shows the trade-off. Anki is more flexible and better if you already have a deck you trust. Word Atlas wins on out-of-the-box content. A paper list is cheap and forces no screen time, but it gives you no scheduling and no audio.
Caveats you should weigh before you commit
This is where a responsible reviewer has to slow down. The project states plainly that its content was generated with generative AI, the illustrations and sentences were produced that way, and the audio uses ElevenLabs voices. Automated checks cover structure and consistency, but the entries haven't been manually proofread one by one. Expect the occasional wrong sense, awkward example sentence, odd illustration, or mispronunciation.
It also isn't an official IELTS word list, and nobody can promise a band score from any app. Use it as a recognition and recall tool, then check unfamiliar or high-stakes words against a learner's dictionary such as the Oxford Advanced Learner's Dictionary or Cambridge Dictionary. If you find an error, the project asks you to report the entry ID, the steps to reproduce it, and the expected result, plus a dictionary reference if you have one.
Practical limits also apply. Running the project locally requires Node.js 20 or later. The repository includes the full images and audio, so the first clone can be large. Progress lives on your device, so export a backup before you switch browsers or phones.
How to use 7,298 words without wasting them
Here's the mistake most candidates make: they treat a 7,000-word list as a to-do list. Vocabulary knowledge is cumulative, and a word you can recognise isn't the same as a word you can use under exam pressure. Your goal is to move words from recognition to production.
A simple routine works well:
- Pick one topic and one CEFR band, such as B2 environment, to keep the set coherent.
- Learn a small batch of new words each day. Around 15 to 20 is a sustainable starting point for most people, though the right number depends on your schedule.
- Write one 100-word paragraph the next day that uses those words, and check each one against its entry's collocations.
- Let FSRS handle the repeats instead of manually re-reviewing.
Most people stop after the first two steps. Don't. The production step is where the score gains usually come from, because it's the part examiners actually see.
Frequently Asked Questions
What is the Word Atlas IELTS vocabulary app?
Word Atlas (词光) is an open-source study app with 7,298 IELTS-oriented word entries, illustrations, UK and US audio, and FSRS spaced review. It runs without an account, stores progress on your device, and is available as a hosted site or as code you can run locally with Node.js 20 or later.
Is there a free IELTS vocabulary app with pictures?
Yes. Word Atlas is free, and every entry includes an original illustration. Its code is MIT licensed and its dataset is CC BY 4.0, so you can also reuse the images and words in your own tools, provided you credit the contributors.
How do I use an IELTS vocabulary app effectively?
Learn a small set of words each day from one topic and CEFR level, then recall them without hints. Write short paragraphs using those words the next day. Recognition is easy, but production is what raises your Lexical Resource score.
Why does spaced repetition beat re-reading word lists?
Spaced repetition schedules each review just before you're likely to forget, and retrieval practice strengthens memory more than re-reading does. FSRS personalises those intervals using your own review history, so hard words come back sooner and easy words wait longer.
Can I trust the definitions and audio in Word Atlas?
Use them as a starting point, not a final authority. The project says its content was generated with AI and hasn't been manually proofread entry by entry, so check unfamiliar or high-stakes words in a learner's dictionary before you rely on them.
Vocabulary gains come from retrieval and use, not from how many words you've opened. If you want to test this, pick 15 words from one topic tonight, write a 100-word paragraph with them tomorrow, and note how many you can recall without a hint. Then share this article with someone who's stuck on their IELTS vocabulary.