italki vs Preply for Learning Japanese: Which Tutors Are Better?

Disclosure: This article contains affiliate links. If you sign up through one of my links I may earn a commission, at no extra cost to you. Nobody from either company has seen this article before you, and neither company has paid for or reviewed anything here. To be straight with you about what this is: I have not taken lessons on both platforms. This comparison is built from the two companies’ published pricing, terms and refund rules, checked directly on their sites in July 2026, plus what I can tell you from Tokyo about what “certified Japanese teacher” actually means here.

Open italki and Preply in two browser tabs and search for a Japanese tutor. You will find one charging $22 an hour and one charging $24, with near-identical profile photos, near-identical blurbs, and in some cases the same person teaching on both. They look interchangeable. Does it matter which one you pick?

The tutors, probably not. The platforms, absolutely — but not for the reason most comparison posts give you. The tutor pool is not really the difference. What’s different is what happens to your money when your life gets in the way: when work travel eats a week, when you get sick on a Tuesday, when you decide in month three that you want to slow down. One of these platforms holds your money loosely. The other one holds it on a 28-day clock.

Buying a lesson package you never finish is close to a rite of passage, and nobody puts it in the guidebooks. So this comparison is built around the boring mechanics — cancellation windows, expiry rules, what the tutor actually takes home — plus one thing no US-based reviewer can tell you, which is what “certified Japanese teacher” now legally means in Japan.

All prices and counts below were checked on July 27, 2026, in USD, from the public listing pages. The yen sat at about ¥163.6 to the dollar that morning, which matters more than you’d think — I’ll come back to it.

Who This Comparison Is NOT For

Save yourself the read if any of these are you.

• If you haven’t started at all. A tutor is not where you begin Japanese. Learn hiragana and katakana first — two weeks, free — because paying $25 an hour to have someone drill あいうえお with you is the single most common way beginners waste money. My broader breakdown of learning Japanese online covers what to stack underneath tutoring.
• If you want a curriculum handed to you. Both platforms are marketplaces, not courses. There’s no syllabus, no progress bar, and nobody chasing you. Some tutors bring structure; most bring whatever you ask for.
• If you need a certificate or academic credit. Neither platform issues one. If you need documented Japanese study hours for a visa or a university application, you need a registered school, not a marketplace.
• If you’re already at N1 level. At that point you want a specialist — a business-keigo coach, an interpreting tutor, a writing editor — and you’ll find them through professional networks or Japanese-side platforms, not by filtering a marketplace by price.
• If $60 a month is your absolute ceiling and it can’t move. Two tutoring sessions a week at real market rates is $180–280 a month. One session a week is $70–140. Below that, apps and self-study give you more per dollar until you can afford to talk to a human regularly.

The Three Things I Actually Compared

1. What happens to money you don’t spend this month. This is the whole ballgame and almost nobody writes about it. Preply runs on a subscription: you pick a number of lessons, and it charges you every 28 days. Per Preply’s own help centre, lessons you haven’t booked expire at renewal and don’t carry over. italki has no subscription at all — you load credits into a wallet and spend them whenever. Both approaches are defensible. Only one of them punishes a bad month.

2. How the tutor gets paid. You are not just buying a lesson, you’re buying a tutor’s motivation to prepare for it. Preply publishes its commission ladder on its own site: it takes 33% from tutors with 0–20 hours taught, then 28% (21–50 hours), 25% (51–200), 22% (201–400), and 18% above 400 hours. It also keeps the entire fee from your first lesson with a new tutor — the tutor earns nothing from that session. italki’s Teacher Policy, by contrast, doesn’t publish a number at all; it just says fees “may vary based on the type of Lesson or Lessons, the number of Lessons in a package, and the Teacher’s overall performance.” In other words: unpublished. Neither arrangement is charity, but a new tutor on Preply keeping 67 cents on the dollar — and nothing on your trial — changes who’s willing to be there in year three.

3. Whether the teacher is actually in Japan, and what their qualification means. For Japanese specifically this is not snobbery, it’s scheduling and slang. It also became a much more concrete question in April 2024, when Japan turned “Japanese language teacher” into a national qualification. More on that below, because it’s the part I can tell you that a reviewer in Ohio can’t.

italki vs Preply for Japanese: The July 2026 Numbers

The Tutor Numbers, and Why the Big One Is Misleading

Preply lists 4,936 Japanese tutors. italki lists 1,184. If you shop on inventory alone, that’s a four-to-one blowout and we can all go home.

Except the sizes measure different things. Preply’s count is its whole historical roster of people who signed up to teach Japanese, including tutors who are not native speakers, tutors with two reviews from 2023, and tutors whose calendars have been empty for months. italki’s number is what it says on the page: tutors available.

Here’s the part that surprised me. I pulled the first page of results on both sites the same morning. Preply’s ten featured tutors ran $22 to $50 an hour, averaging about $35 — that’s from the page’s own structured data, on a page that markets a “$19 average” and “from $3.” italki’s first page of twenty teachers ran $15 to $55, averaging about $29. Seventeen of those twenty were listed as Japanese nationals; the other three were Chinese.

So “italki is the cheap one” doesn’t survive contact with the actual listings. Both platforms front-load their expensive, well-reviewed teachers, because that’s what converts. The real price gap appears only when you start filtering — and that’s where italki’s structure matters.

The Community Tutor Tier Is the Actual Difference

italki splits its teachers into two categories: professional teachers, who have to submit teaching credentials, and community tutors, who are native or fluent speakers without formal qualifications. Per italki’s own pricing guide, community tutors run roughly $4–20 a lesson and professional teachers $10–40.

Preply doesn’t make that split. Every listing is a “tutor,” and the profile text is all you have to work out which kind you’re hiring: a trained teacher, or a friendly student.

For Japanese, that distinction saves real money, because you need two different things at two different prices. You need someone who can explain why は and が split a sentence the way they do — that’s worth $30 an hour and you might need it once a week. And you need vastly more hours of somebody patiently letting you fumble through describing your weekend — which is worth $8, and you should be doing it three times a week.

Buying the second thing at professional-teacher prices is how people burn $300 a month and quit in October. Two community-tutor sessions a week at around $9 each costs less per month than four professional lessons, and it is the half of the equation that actually builds fluency. It is also where the vocabulary lives that textbooks skip: no beginner textbook I’ve seen in a Japanese bookshop contains 「やばい」, and no $35-an-hour structured lesson is going to spend twenty minutes on it either.

See what Japanese teachers on italki charge right now — filter to community tutors and sort by price, and the picture changes completely from the front page.

The Money Mechanics: Where People Get Burned

This is the section I’d read twice.

Preply’s 28-day clock. You subscribe to a number of lessons per cycle — minimum four — at your tutor’s rate at the time you signed up. Cycles are 28 days, not calendar months, so your billing date walks backwards through the month and eventually you’re being charged thirteen times a year, not twelve. Lessons you booked are safe. Lessons you didn’t book expire at renewal. Extra lessons you buy as top-ups do roll over, for up to two cycles. To stop a charge you have to cancel at least a day before the billing date; cancelling on the day doesn’t stop it. A refund is a separate request, generally available if you cancel with no lessons used in that cycle and ask within 28 days of the charge.

None of this is hidden — it’s all in Preply’s help centre, plainly written. But it means a two-week work trip costs you money, and the fix, remembering to pause before the billing date, is exactly the admin that busy people forget.

italki’s wallet. You buy credits, credits sit in your wallet, you book lessons whenever. Purchases are final, so what goes in can’t come back out to your card — that’s the genuine downside and I’d load $50 at a time rather than $300. Credits expire after twelve months of account inactivity, with one reactivation possible if they’ve been expired under a year. Cancel a lesson more than 24 hours ahead and 100% of the credits go back to your wallet. If a teacher declines or lets a request expire, you get a full wallet refund automatically. After a lesson, either side has 72 hours to hit “lesson incomplete” and propose a refund.

The one place Preply is genuinely more forgiving: the cancellation window. Preply gives you back a lesson if you cancel 12 hours out. italki wants 24. If your life falls apart at 6pm for a 9am lesson, Preply covers you and italki doesn’t. That’s a real advantage and I’m not going to pretend otherwise — it’s just smaller, to me, than the expiry rules on the other side of the ledger.

Trial Lessons: The Most Revealing Difference

On italki, a trial is 30 minutes, priced by the teacher with a $5 floor, usually 30–50% off their standard rate, and you get up to three of them across different teachers. Budget $15–25 to audition three candidates. That’s a rounding error against a year of lessons, and italki says it doesn’t take commission on trials, so the teacher is being paid for their time.

On Preply, your first lesson with a tutor is a trial at their normal rate, and Preply keeps the entire fee. The tutor teaches you for free. If it goes badly — wrong style, bad connection, tutor no-show — you can request a free replacement, but the account cap is two replacements total, and a “replacement” just returns the lesson to your balance rather than granting you a bonus one. After two, you pay for every new trial.

Think about what that does to incentives. A tutor doing five trials a week on Preply is donating five hours to the platform, so the pressure is to convert you into a subscription fast. A tutor doing five trials on italki got paid for five half-hours. This is a tendency, not a law — plenty of Preply trials are excellent and plenty of italki trials are a wasted half-hour. But it explains a pattern learners report constantly: Preply trials that feel like sales calls.

What This Looks Like From Japan

Now the part I can add that most reviews can’t.

“Certified” now means something specific here. On April 1, 2024, Japan introduced a national qualification for Japanese language teachers — 登録日本語教員 (tōroku nihongo kyōin, “registered Japanese language teacher”) — with a state exam attached. In the first sitting, 17,655 people took it and 11,051 passed, a 62.6% pass rate. The second exam, on November 2, 2025, drew 17,597 candidates with 11,876 passing, 67.5%. Before this, the closest thing to a standard was the 日本語教育能力検定試験 or a 420-hour teacher training course, which typically runs ¥400,000–600,000 out of pocket here.

Practical upshot: when a teacher’s profile says “certified,” you can now ask which certification. A tutor holding the national registration, the older 検定 pass, or a 420-hour course certificate has genuinely studied how to teach this language to foreigners — how to sequence particles, when to introduce 敬語, why your textbook’s 〜てしまう explanation is confusing you. A native speaker without any of that is a wonderful conversation partner and not a teacher, which is fine as long as you’re paying conversation-partner prices. italki’s two-tier structure makes that distinction visible on the listing page. On Preply you have to dig it out of the profile text yourself.

Time zones work in your favour, if you book Japan-side teachers. Japan is UTC+9 with no daylight saving, so the arithmetic is stable all year. A teacher’s 9pm Tokyo slot — the most common evening availability here — lands at 8am in New York, 5am in Los Angeles, 1pm in London, and 10pm in Sydney. Their 7am slot lands at 6pm the previous day on the US East Coast, 3pm on the West Coast. East Coast learners have it best: a Japan-based teacher’s early morning is your after-work hour, and those slots are usually wide open because most of their students are European. Scan a Japan-side teacher’s calendar around 7:30am Tokyo time and you will typically find it empty.

The weak yen is doing you a favour, and the sticker price hides it. At ¥163.6 to the dollar on the morning I checked, a teacher in Japan who prices at ¥4,000 an hour — a completely normal domestic private-lesson rate — is charging you about $24.50. Five years ago that same lesson was noticeably more expensive in dollars. Japan-based teachers tend to anchor their pricing to what feels reasonable in yen, which quietly makes them the value pick on both platforms.

The domestic option nobody outside Japan mentions. Cafetalk is a Japan-based lesson marketplace most English-speaking learners have never heard of. It runs on points — 1 point equals ¥1.1 including tax — and Japanese lessons commonly sit at 1,500–3,900 points per hour, roughly ¥1,650–4,290, or about $10–26. The interface is a bit 2012 and English support is thinner, but the teacher pool is heavily domestic. I mention it because “italki or Preply” isn’t the whole universe.

And on materials: if a tutor asks you to buy みんなの日本語 (Minna no Nihongo), the standard beginner textbook in Japanese classrooms here, the publisher’s list price for Book I is ¥2,750 — about $17. Japanese-side teachers usually teach from it or from Genki, and it’s worth asking which before your first paid lesson, because working through a shared textbook is what turns a chat into a course.

italki: What I Like and What Annoys Me

What works. The pricing model fits how adults actually learn: irregularly, in bursts, with gaps. Credits sit still when you’re busy. The community-tutor tier lets you buy cheap speaking volume, which is the bottleneck for nearly everyone. Cheap 30-minute trials mean auditioning three teachers costs less than a bar tab. The Japanese teacher pool skews genuinely Japan-resident, so the time zones and the slang both work out. And nobody charges you in a month where you take zero lessons.

What annoys me. The platform feels a half-generation behind — scheduling, messaging and the classroom are all clunkier than Preply’s, and the mobile app is worse than the website. Credits are one-way: money in is money committed, so don’t front-load a year’s budget. The 24-hour cancellation window is stricter than Preply’s 12. Community-tutor quality is a lottery, and reading reviews carefully is not optional. And italki not publishing its commission rate is a small transparency ding, even if it’s below Preply’s opening 33%.

Who it’s for: anyone whose schedule is unpredictable, anyone who needs a high volume of cheap speaking practice, and anyone playing a multi-year game with Japanese rather than a three-month sprint.

Preply: What It Does Better

I want to be fair here, because Preply does two things better and one of them might be decisive for you.

The product is more polished. Booking, rescheduling, the classroom, the progress tracking, the vocabulary tools — it’s a smoother piece of software, and if you’ve ever abandoned something because the UI made you tired, that’s worth money.

The subscription is a commitment device. This is the strongest argument for Preply and the reviews that mock it are missing the point. Some people learn a language only when money is already spent and a lesson is already on the calendar. If you’re one of them, being billed every 28 days for four lessons is a feature, not a trap: cancelling feels worse than showing up, and that discomfort does the scheduling for you.

The 12-hour cancellation window is friendlier than italki’s 24, as noted.

Where it falls down for Japanese specifically. The tutor economics are hard to ignore: 33% off the top for new tutors, 100% of your first lesson, and a ladder that only reaches 18% after 400 hours taught. That pushes experienced Japanese teachers toward platforms and private arrangements where they keep more, and it makes trials feel transactional. The expiry of unbooked lessons at renewal punishes exactly the months when life is chaotic. And without a community-tutor tier, cheap conversation volume is harder to assemble — you’re mostly choosing between “professional rates” and “unknown quantity.”

Who it’s for: learners who know, honestly, that they will not show up unless they’ve prepaid and committed, and who want a weekly slot with one tutor rather than a flexible mix.

What I’d Actually Do, by Situation

You’re a beginner who wants to speak, on $80–120 a month. italki, two community-tutor sessions a week at $8–12 each, plus one professional teacher session a month for grammar cleanup. That’s roughly $90–110 and it’s the highest-return tutoring spend I know of for Japanese. Run it alongside a structured course rather than instead of one — the full stack comparison has my pick for the course layer.

You have a hard weekly slot and a real risk of quitting. Preply, four 50-minute lessons per cycle with one tutor. Set a calendar reminder two days before each billing date so you’re never surprised, and book your lessons the moment the cycle opens so nothing expires.

You travel constantly or your work hours are unpredictable. italki, no contest. Load $50 in credits, book 48 hours ahead when you know your week, and accept the 24-hour cancellation rule as the price of never paying for a month you missed.

You’re prepping for the JLPT in December. A professional teacher on italki with stated JLPT experience, once a week from August, going to twice a week in October. Ask before booking whether they teach from 新完全マスター or 総まとめ — the answer tells you a lot about how systematic they are.

You want to try both. Cheap enough: three italki trials at roughly $5–8 each plus one Preply trial at full rate is $45–70 total, and two weeks later you’ll know which rhythm you’ll actually keep.

The Verdict

Preply builds better software. italki has the better deal for learning Japanese, and it isn’t close once you look past the front page.

The reasons stack up: no subscription clock, so a bad month costs you nothing. Credits that wait for you instead of expiring at renewal. A community-tutor tier that lets you buy the cheap speaking volume that actually moves your Japanese, at a third of professional rates. Trials at $5 for thirty minutes, with the teacher getting paid for their time so the session is a lesson rather than a pitch. And a Japanese teacher pool that skews properly Japan-resident, which is what makes the time zones and the 「やばい」 both work out.

Preply’s cancellation window is better and its product is nicer to use. If you know that only a recurring charge will make you show up, take it — that’s a real thing about yourself and worth respecting. Everyone else: put your money where it doesn’t evaporate.

If you’re starting this week, do the cheap version first. Book three 30-minute trials with three different teachers — one professional, two community tutors — on three different days, and give each one the same opening: “I want to talk about my weekend for twenty minutes, please correct me when it matters.” You’ll know by Sunday who you want to see again.

Browse Japanese teachers on italki and book a 30-minute trial

All prices, tutor counts and policy details above were checked on July 27, 2026, from italki’s and Preply’s own listing pages, help centres and published terms. Both platforms change pricing and policies without much warning, so if you’re reading this months from now, verify the numbers before you commit.