Duolingo
Designing seamless writing experiences for non-Latin language learners
Problem
Current lessons teach recognition and pronunciation but do not support in-app handwriting. This forces learners to rely on external resources like YouTube or tracing books, limiting confidence and fluency when writing on their own.
Goal
Integrate handwriting practice for non-Latin languages directly into Duolingo’s lesson structure, allowing learners to practice tracing and writing—not just reading and listening.
Outcome
A redesigned lesson flow that focuses on tracing exercises, stroke order guidance, audio support, and real-time feedback—helping learners build handwriting skills alongside vocabulary and pronunciation.
Background
Duolingo is one of the most popular language-learning platforms today, offering over 40 languages and reaching a global audience—80% of its users are outside the US. With ~116.7 million monthly active users and $531 million in revenue (2023), Duolingo appeals to a broad demographic, skewing slightly younger (18–34) but serving learners of all ages, whether for K–12 education or professional development. Its gamified approach makes learning feel motivating, fun, and accessible.
As someone with a 2,000-day Duolingo streak, I’ve experienced both its strengths and gaps firsthand. When I set out to learn Arabic, I quickly realized that while the app helped me memorize characters, it lacked meaningful ways to practice handwriting. This missing hands-on interaction dampened my motivation, leading me to switch to a Latin-based language that felt easier to reinforce offline. Although Duolingo later introduced limited writing practice, it remains separate from the main lesson flow. For this project, I set out to design a more seamless and engaging solution—embedding handwriting practice directly into the core lesson experience to help learners feel supported, confident, and connected to the language they’re studying.
OBSERVING, LISTENING, LEARNING
SWOT Analysis
To better understand how other language‑learning platforms approach handwriting, I conducted a SWOT analysis of three direct competitors: Babbel, Mango Languages, and Rosetta Stone. Through this analysis, I was struck by how advanced and varied their speaking and reading tools were—each offered creative ways to engage users in verbal practice. Yet despite this strength, none provided integrated handwriting instruction for non‑Latin scripts, highlighting a significant gap in supporting learners who want to write as well as speak and read.
The key gap: While these platforms support speaking and reading, they don’t help users actively practice handwriting—leaving learners to rely on memorizing characters instead of truly learning to write them.
Babbel
Rosetta Stone
Mango Learning
Research Interviews
I began by identifying interview participants who were either fluent in a non-Latin language or currently learning one. In total, I conducted five interviews: two with active Duolingo users, two with multilingual speakers from birth, and one with someone who became fluent through formal study.
A total of 14 questions were asked. A sample of these questions being:
How do you maintain your knowledge of a language?
When learning a language, what techniques do you find the most impactful?
Please rate your confidence in a few components of language (reading, speaking, and writing) on a scale of 1 to 10.
Once I completed the interviews and created an affinity map, key themes emerged:
Even advanced learners lacked confidence in writing by hand.
Fluent or semi-fluent writers improved through hands-on practice and repetition
Participants who attempted to learn how to write on their own, but had no real-time feedback were unable to know if they were writing correctly.
Only the participant who studied formally rated their writing ability with confidence, offering insight into effective handwriting practice. This highlighted that focused, repetitive writing exercises are essential for building deeper understanding and connection to the language.
To better understand the core challenge, I analyzed findings from my research, interviews, and competitor analysis, which revealed the following key insights:
UNCOVERING THE CORE CHALLENGE
Learners feel confident recognizing and typing characters, but not writing by hand.
Recognition alone doesn’t translate into handwriting fluency, leaving learners unsure if they’re doing it correctly.Lack of integrated writing practice disrupts learning flow.
Users often rely on external tools (YouTube, tracing books), which interrupts momentum and feels disjointed from the main lesson structure.Feedback and a sense of progress instill confidence
Without guided practice and real-time correction, learners can’t see their progress, leading to hesitation and lower motivation.
Motivation is personal—but unsupported.
Each language learner had an emotional reasons (family, history, travel) driving them to learn or sustain a language, but Duolingo doesn’t fully support these meaningful goals.Competitors also neglect balanced skill-building.
Other language apps focus heavily on reading and speaking, while handwriting remains an afterthought or is completely absent.
These insights led to a clear UX problem statement that framed the design challenge:
Duolingo users lack integrated handwriting practice for non-Latin scripts, forcing them to rely on external tools—weakening confidence in real writing and increasing the risk of disengagement from the platform.
To bring this problem to life and better understand user needs, I developed the persona of Mina, a late-20s Korean adoptee who wants to write a letter to her birth mother in Korean. Mina’s motivation is deeply personal, but her journey is slowed by unclear guidance and scattered practice resources. Building this persona helped me frame the user’s emotional drivers—connection, identity, and confidence—and map out where current gaps exist in the experience.
By defining Mina’s story, needs, and frustrations, I was able to validate why adding handwriting practice directly into Duolingo’s lesson flow could meaningfully improve user confidence and retention—and why this challenge became central to the next design phase.
Framing Mina’s challenges and motivations led to a focused question:
How might we help learners feel guided and supported when writing,
rather than relying on external tools?
From there, it allowed me to start to prioritizing the features that would need to be added to the existing Duolingo structure, which included:
Updated touch interface that allows users to trace. The interface would need to register stroke location within the touchbox to determine if the character had been written correctly.
Outlines of characters that would allow users to trace within dotted/dashed areas during lessons.
Audio pronunciation alongside writing practice tracing (although this is unable to be presented in the prototype due to technical limitations)
The process began by mapping out a user flow that shows where handwriting practice could naturally fit within existing lessons.
BUILDING THE EXPERIENCE
From there, I began sketching low‑fidelity wireframes on paper. The goal at this stage was to map out ideas and identify how handwriting could feel integrated rather than isolated. Once these concepts were outlined, I moved into digital wireframes to refine layout and structure.
FROM SKETCHES TO SCREENS
Mid‑fidelity prototypes focused on refining visual details, spacing, and the way users could view and correct mistakes. Throughout the process, the goal was to balance guidance with exploration—supporting users without adding unnecessary friction. The task flow shown below guides users through learning to write the Korean letter ae: first by tracing each stroke, then by writing it from memory.
I conducted user testing at this stage with five new users, starting them at a recreation of the Duolingo home screen and having them complete the above lesson. While all testers were able to complete the task, there was was notable friction at the following points:
Instructions for open-ended writing tasks needed to be clearer.
Learners wanted more repetition and guided tracing before attempting to write characters from memory.
Some buttons, like “Practice Again,” were overlooked or misunderstood.
Based on this feedback, the following updates were made moving into high fidelity:
When writing from memory instructions and a pen indicator to guide the user
After
Before
Added additional tracing practice in which users trace the characters without the guide arrows before writing from memory.
Original “Continue” screen separated into two parts: one screen offering users the choice to keep practicing or continue, and a second screen that indicates the next practice activity.
Before
After
After
COHESIVE BRANDING
Although the majority of Duolingo’s existing branding was used, I adapted and expanded it to fit the new handwriting feature.
To allow for better flow between stages of the practice I added prompting cards, utilizing existing Duolingo graphics.
To show the characters in context, I created Duolingo-esque cartoons.
A SEAMLESS INTEGRATION
As I developed the high-fidelity prototype new elements were added. Along with the addition of a tracing exercise, I added an additional character for the user to practice, expanding this practice into a comprehensive lesson. Duolingo’s branding was integrated into the design, smoothing transitions between practice steps.
Tracing with stroke direction indicators
Tracing without stroke direction indicators
Writing from memory
Second letter “gye” follows same sequence
I then conducted a second round of usability testing with 5 users, having them complete a task flow. Users were to navigate from the Duolingo homepage, through the “Learn Korean” lesson, and return back to the homepage. Success was measured by task completion time (under six minutes) and the number of flow divergences, which were flagged by an “incorrect” popup. All participants completed the task within the time limit and without divergence. Alongside positive feedback, users also raised questions and encountered minor points of friction, leading to the following key insights:
Users indicated that learning two words is manageable and not too overwhelming.
Providing an example of the word in context would help users understand its usage among other characters.
Requiring users to click on the pen to start writing from memory slows down the process and is not immediately intuitive. It was suggested that writing should be triggered by clicking into the open writing box.
Based on this feedback, the following updates were made:
Lesson titles and section names updated showing the lesson is specifically a writing lesson.
Text stating that the order of the strokes is important.
Clearer and more supportive error messaging.
Before
After
Removed the need to touch the pen when writing. It now auto select the pen when a user touches the writing box.
Before
After
After writing the letter from memory, users see a context screen showing the character in use, complete with phonetics, visuals, and audio.
LESSONS FROM THE PROCESS
In EdTech, user feedback guides the way
Determining how much repetition learners needed for each character wasn’t obvious at first. Research from learning science offered a starting point, but observing users directly proved invaluable. Watching how users interacted with tracing, and noting both their verbal comments and nonverbal cues, helped pinpoint the right balance between guidance and challenge. This user-centered approach ensured the final design felt supportive rather than repetitive.
Self-editing keeps design aligned with the goal
At times during the process, I added extra exercises—like additional characters to trace or matching tasks. After fully designing these features, I realized they didn’t fit the original goal and conflicted with user feedback, costing time and effort. This experience taught me that while exploring beyond the initial scope can spark ideas, it’s essential to pause and reflect before investing deeply in features that may not truly serve users.