Loading

Persian as a Three-Layer Cake: Apps, Media, and Humans

Asking "which app is best for Persian" is the wrong question.

The right question is: what three layers do you stack, and in what ratio, given that no single layer reaches even halfway on its own. Persian is one of the most under-served languages in commercial language tech. There are roughly 110 million speakers across Iran, Afghanistan, Tajikistan, and a global diaspora that is denser per capita than the Italian one. There are maybe four serious commercial apps. That mismatch is the structural fact that should shape every choice you make.

If you walk in expecting a Duolingo-clone to carry you to B1, you will quit by month four. If you walk in expecting to stitch together a small commercial layer, a rich free Iranian media layer, and a tutor layer, the language opens up faster than you would think.

I am Alej Pascual. I speak 11 languages, Persian is one of them at B1, and I learned it because I had Iranian friends in Mexico City and London and went to Tehran twice in the years before sanctions tightened. The rest of my polyglot credentials live on my LinkedIn. What follows is the layered map I wish someone had handed me in 2019.

The three layers, in order:

Layer 1: the four commercial apps worth your money

The Persian commercial app market is small, which is a gift in disguise: you do not have to triage twenty options. Four apps cover the entire credible commercial layer for Persian as of 2026. If you have used something else and it worked for you, keep it. If you are starting today, these are the four to consider.

Chai and Conversation is, honestly, the single best Persian resource that exists in any language. Leyla Shams runs it. Tehran-flavored Iranian Farsi with a Los Angeles diaspora sensibility. Free podcast archive covers absolute beginner through advanced. Premium membership at around 15-20 dollars a month adds PDFs and lesson plans. The pacing is the rare mix of warm and rigorous. Script handling is minimal, so pair it with one of the others below for reading.

Pimsleur is the only major commercial brand that ships Persian, and they ship two separate courses: one for Iranian Farsi, one for Dari. The Dari course is unusually thoughtful given how niche the market is. Audio drills only, zero script, caps at A2. Strong for the first three to six months because the consonant work is genuinely accurate.

Mynago is mine, so weigh it accordingly. I built the Persian course to generate daily lessons in Iranian Farsi by default, with scenarios drawn from actual Tehran life: bazaar haggling, the elaborate ta'arof politeness rituals, ordering chelo kabab, navigating a shared taxi. Script is taught in the same lessons as audio, not bolted on. Heritage learners switch the variant flag at onboarding so the example sentences align with Dari instead. The slot it owns: structured daily input with the script and the audio integrated rather than separated.

Anki with the community "Farsi 5000" frequency deck, built from Iranian written corpora. Free on Android and desktop, 25 dollars one-time on iOS. Twenty minutes a day. The first thousand words cover a startling percentage of daily Persian speech. Ugly interface, unbeatable retention math.

What does not exist in this layer: a Duolingo course (no Persian on Duolingo), a Babbel course (no), a Rosetta Stone course (sometimes-not-really), a Busuu course (no). The market is genuinely thin. That is why the next two layers matter.

Layer 2: the Iranian and Afghan media wiring that costs nothing

This is the layer most "best Persian apps" articles skip entirely, and it is the one that does the heaviest lifting from intermediate onwards.

Persian-language media is enormous and largely free. The Iranian diaspora produces an absurd amount of YouTube, podcast, Telegram, and Instagram content. The Iranian state broadcaster, the BBC Persian service, Manoto, Iran International, and dozens of independent diaspora channels publish daily. Afghan media in Dari runs out of Kabul (under constraints), Pakistan, Iran, and the European Afghan diaspora. None of it is gated behind a subscription. Almost none of it is indexed in English-language app lists.

The tools that turn this firehose into a study system:

Forvo is the pronunciation dictionary I check more than any other. Crowd-sourced native recordings filterable by speaker country, which means you can compare an Iranian and an Afghan saying the same word and hear the gap. Free.

LingQ is the import tool. You paste in a transcript or feed in a YouTube link with Persian subtitles, LingQ tracks what you know and what you do not, and you read through with click-to-define. The Persian library on LingQ is thinner than the French one but the import feature is the unlock. A 30-minute Manoto interview or a BBC Persian news segment becomes a week of study material.

BBC Persian and Iran International run their daily news bulletins on YouTube with reliable subtitles. The accent is Tehran-leaning standard. The vocabulary is current. The clip lengths are 60 to 90 seconds, which is the right size for early intermediate input.

Manoto is the Iranian diaspora satellite channel, free on YouTube. Talk shows, cooking, music. Less formal than the news, more representative of how people actually speak.

Telegram channels. This is the Iranian internet's home court. Persian-language Telegram has news, comedy, language explainers, graded reading material, and meme channels. Finding them requires Persian-script search, which loops back to the script-first rule. Once you can read the channel name, the open Iranian internet is the immersion library no app can replicate.

For Dari specifically. Afghan-diaspora YouTube has expanded sharply since 2021. Kabul-based broadcasters that survived the political upheaval still publish. The DLI (US Defense Language Institute) and Indiana University Persian programs have made Dari course PDFs and audio freely available, which is the highest-accuracy free academic material on the open internet.

For Tajiki. The Indiana CeLCAR center maintains the most credible free Tajiki materials. The US Peace Corps Tajik training manual is freely downloadable. The Aga Khan University Tajiki language program is a third option a Dushanbe reader pointed me to. Tajiki commercial apps are essentially non-existent; this layer carries almost the whole weight.

The ratio of this layer to layer 1 changes with level. At A1 you are barely touching it. By B1 it is half your study time. By B2 it has eaten almost all of it.

Layer 3: the human tutor who turns the other two into speech

Persian reads as a tonal-free Indo-European language with reasonably forgiving grammar. The vowel system and the consonants behind it are where the wheels come off for English speakers. No app on this list will fix that. A tutor will, in twelve sessions.

Italki is the marketplace. Persian tutor supply is healthy: hundreds of Iranian tutors active, dozens of Afghan tutors for Dari, a handful for Tajiki. Community tutor rates run 8 to 15 dollars an hour. Professional teachers run 20 to 35 dollars. For the first three months, a weekly thirty-minute session is enough. By month six, twice a week if you can afford it.

Preply is the alternate marketplace. Slightly different tutor pool. Compare both before you settle on a teacher.

The five sounds your tutor will hand-correct. I will not pretend an app will teach you these. A tutor will. The "kh" (خ), the "gh" (غ/ق), long-vs-short vowels (ā vs. a, ī vs. i, ū vs. u), word stress (final-syllable in nouns, fixed under suffixation), and the izafe particle ("-e" or "-ye") that links nouns to adjectives. Apps gloss every one of these. Pimsleur drills the izafe, half-decently. Everything else needs human ears.

The Tehran-vs-LA-diaspora register question. A tutor in Tehran will calibrate you toward the Iran-internal register. A tutor in Los Angeles, Toronto, or Hamburg will calibrate you toward the diaspora register. The two have diverged over forty years in ways that matter for vocabulary. Pick based on who you intend to talk to.

The conversation circles. In cities with a meaningful Iranian diaspora (LA, Toronto, London, Hamburg, Sydney), Persian conversation circles run at cultural centers and university programs. Free. They are the analog of layer 3 if your budget cannot stretch to Italki. Search "Persian conversation Vancouver" or "Farsi exchange Berlin" and you will find them.

The ratio that actually works (and how it changes by month)

This is the part most learners get wrong. They run one layer hard, hit a wall, then quit. The ratio matters as much as the components.

Months 1 to 3. Layer 1 carries 70 percent of your time. Layer 2 sits at 10 percent (Forvo plus a daily 60-second BBC Persian clip you will only half-understand). Layer 3 is the remaining 20 percent (one weekly tutor session, even if you can barely say anything). The script work happens here. Do not skip it.

Months 4 to 6. Layer 1 drops to 40 percent. Layer 2 climbs to 35 percent (Chai and Conversation premium archive, a single Manoto episode per week, daily news clips). Layer 3 holds at 25 percent (one tutor session a week, possibly two).

Months 7 to 12. Layer 1 drops to 15 percent (Mynago daily, Anki maintenance). Layer 2 reaches 50 percent (LingQ-imported real content, podcasts in the background, books). Layer 3 holds at 35 percent (two tutor sessions a week, with one focused on reading aloud and the other on free conversation).

Year 2 onwards. Layer 1 is essentially Anki only. Layer 2 is your main input. Layer 3 stays. The apps that scaffolded you have done their job.

If any of this sounds like a lot of unsexy book-keeping, that is because the people who get to B2 in Persian without living in the country all run something like this. The people who try to find "the one best app" stall at A2 and quit.

Dialect picker: Iranian Farsi, Dari, or Tajiki

A small section because the question comes up constantly.

Default to Iranian Farsi. Roughly 60-70 million speakers vs. 20-25 million Dari and 8-10 million Tajiki. The largest body of recorded media. Almost all classical poetry (Hafez, Saadi, Ferdowsi, much of Rumi) was composed in what evolved into Iranian Persian and is read today by speakers of all three variants. The deepest stack of teaching resources. If you are unsure, start here.

Switch to Dari if you have Afghan family, an Afghan partner, or specific work in Afghanistan or with the Afghan diaspora. The vocabulary and pronunciation differences will matter from day one. Pimsleur ships a dedicated Dari course; Mynago supports a Dari variant; the DLI and Indiana University materials cover the academic foundation.

Switch to Tajiki only if you specifically need to live, work, or marry in Tajikistan. The Cyrillic script is a real time investment that does not transfer to the rest of the Persian-speaking world. The commercial layer is almost empty; layer 2 (Indiana CeLCAR, Peace Corps, Aga Khan University materials) does most of the work.

For mixed-heritage unsure learners. Iranian-Afghan diaspora families are common. Start with Iranian Farsi for three months because the resources are denser, then layer Dari listening on top once the ear is trained. Grammar is identical. You are retraining vowels and swapping a small everyday vocabulary set.

If you want a structured starting line, take the free Persian level test and let the system set up a study plan. It is calibrated against Iranian Farsi by default, with a heritage flag for Dari learners.

The framing I would have wanted in 2019: Persian is not a one-app language. It is a three-layer language. Pick your dialect, set up the three layers in the right ratio, and the language opens up. Get it wrong, and no amount of streak-keeping will save you. Get it right, and you will speak workable Persian by year two even if you never set foot in Tehran.

Update, May 2026: heritage learners, the Telegram boom, and Tajiki additions

Three patterns from reader replies that the original post needed.

Heritage learners run a different ratio. A common pattern: they speak conversational Persian at home, cannot read or write Farsi script, and want to repair the literacy gap. For them, layer 1 looks different (skip the conversational A1 apps, start with a Persian script focused course like the University of Texas Center for Open Educational Resources free materials). Layer 2 stays the same. Layer 3 is sometimes optional because they have family conversations as the human-in-the-loop substitute. I wrote about this more in reclaiming Persian as a heritage learner.

The Telegram and YouTube layer-2 boom is genuinely new. Since 2022 the diasporic Persian content economy on Telegram and YouTube has expanded enough that you can find native-produced vlogs, comedy podcasts, language explainers, and graded reading material almost for free. The catch is the same: finding them requires Persian-script search, which loops back to script-first. Once you can read titles, layer 2 effectively becomes infinite. I underweighted this in the original post.

Tajiki learners pointed me to better layer-2 resources. A Dushanbe reader pointed out the Aga Khan University materials, currently the best free Tajiki structured course online. If Tajiki is your goal, supplement Pimsleur Tajik (where available) with the Aga Khan materials and skip the Iranian-Farsi-translated apps entirely. The lexical and pronunciation differences are larger than apps flag.

The three-layer cake still holds. The apps are the substrate. The media is the immersion. The tutor is the correction. None of them works alone.