Iro AIIro AI Blog
Wondering is a YC-backed app that turns any topic into a path of three-minute lessons. Iro AI is built to make you good at using AI specifically. They overlap enough to be confused and differ enough that the choice is usually obvious once you name what you actually want.
Editorial note: Iro AI is our app. Where it appears below we say so, and it has to earn its spot by the same criteria as everything else.
Want to learn a bit of everything? Wondering. Want to get genuinely good at using AI? Iro.
Wondering is a general learning app: name a topic, get a personalised path of short lessons. Iro AI is a single-subject app: it teaches you to use AI tools well, and it makes you practise rather than read.
We make Iro, so weigh this accordingly. What follows is what each one is actually built for, including the cases where Wondering is the better buy.
Wondering came out of Y Combinator's Summer 2026 batch, built by a two-person team in San Francisco. One of the founders previously worked on NotebookLM at Google, which shows in the product: it is good at turning source material into something structured.
The pitch is Duolingo for anything. You give it a topic, a link, a PDF or just a description of what you want to know, and it builds a path of roughly three-minute lessons with visuals, generated podcasts and exercises. The path adapts to your background, how you prefer to learn, and how deep you want to go.
Building a course from your own material is the headline idea and it works well. Much of the rest of the format, illustrated lessons and audio versions of them, is closer to industry standard than to a moat. The next section covers where the two actually part company.
It is in early access on iOS and web. There is no Android app.
Iro AI is narrow on purpose. It teaches one subject, using AI, and goes deep on it: 29 learning paths, 477 lessons and 2,700+ hands-on exercises across ChatGPT, Claude, Gemini, Perplexity, prompt engineering, agents, automation, vibecoding and AI for specific jobs.
The mechanic is different too. Iro is built around doing the thing and being marked on it. You write real prompts in the Prompt Lab and get them graded, judge whether a model's output is good, spot hallucinations, and duel other learners on timed questions. Streaks, XP and ranks hold the daily habit together, and a visible skill profile shows which of six areas you are weakest in.
It runs on iOS and in any browser, with Android in development.
Iro AI turns ideas like the ones in this post into 5-minute exercises with feedback. Free tier, Pro from $4.17 a month ($49.99 a year, 7-day free trial).
The obvious framing is breadth versus depth, and that part is right. The overlap underneath it is larger than either landing page suggests, which is what most comparisons get wrong.
Both build a course from what you give them. Iro's Custom Paths takes a topic you type, or a TikTok, Instagram or YouTube link, or a screenshot, or a PDF, and returns a full path with lessons, a quiz and a recap.
Both present lessons visually, and both have audio versions of lessons, with Iro's shipping shortly rather than being live today.
Both gamify, which the next section breaks down.
So the format is close to a wash. What is left is the part that is hard to copy: what the app knows, and whether it marks your work.
This is the difference that matters most and the one that is hardest to see from a landing page.
A well-made three-minute lesson with clean visuals feels like learning, and some of it is. You will retain more than from a video you half-watched. But the activity is still mostly consumption, and consumption produces recognition rather than ability. You will know what a system prompt is. Writing a good one under time pressure is a separate skill.
Iro's bet is that the thing you get marked on is the thing you get good at. That is why the exercises are prompt writing scored by a real model, output critique and hallucination spotting, rather than multiple choice about definitions. Generated exercises can test whether you followed a lesson; they are a different thing from a Prompt Lab that reads the prompt you wrote and tells you what is weak about it.
The fair counterpoint: for a lot of subjects, understanding is the whole goal. If you want to grasp quantum mechanics well enough to follow a conversation, a beautifully explained lesson is exactly right and graded drills would be overkill. Practice earns its keep when you intend to perform the skill, and using AI at work is a performance.
Both apps gamify, and it is worth being precise about how, because the marketing shorthand of "Duolingo for X" hides real differences.
Wondering has streaks, personal ones plus Friend Streaks with up to three mutual friends, where you see each other's daily progress and can send a nudge. That is a genuinely good social hook and Iro does not have an equivalent. It also tracks accuracy and activity history and saves Knowledge Cards of key ideas.
Where they diverge is the progression and competition layer. Iro adds XP, six competitive ranks, live ELO-matched duels against other learners on timed questions, daily challenges, chests along the path map, a mascot that evolves as you rank up, and an Operator Build profile that scores six skill areas and names your weakest.
So the honest summary is not that one is gamified and the other is not. Wondering's game layer is built around consistency; Iro's is built around competition and visible progression. If a streak is enough to keep you coming back, Wondering's is a good one, and its friend version is better than anything we ship. If you need a ladder to climb and someone to beat, that is the deeper of the two systems.
Both are free to start. The paid tiers are where they separate.
Wondering Pro is $7.50 a month on annual billing, which is $90 a year. Iro Pro is $49.99 a year, about $4.17 a month, with a 7-day free trial, or $9.99 monthly.
Annual against annual, that is $40 a year more for Wondering, or about 80% more per month.
Two things make the gap wider than the headline. The $7.50 is an early-supporter founding rate rather than the standing price, so new subscribers should check what they are actually quoted. And the free tiers differ in kind: Iro opens every one of its 29 paths on the free tier, paced by a battery that gives you a full lesson a day, so the paid upgrade buys speed rather than access.
Cheaper is not automatically better. But if you are weighing two apps and one costs nearly twice as much for a broader, shallower treatment of your actual subject, that belongs in the decision.
Choose Wondering if your curiosity is broad, you want one app for whatever you are into this month, you learn well from visuals and audio, or you have specific source material you want turned into a course. Also choose it if you are on iOS or web and do not mind an early-access product changing under you.
Choose Iro if the goal is to actually get good at using AI, you want to be marked rather than told, you want a curriculum that goes deeper than an overview on prompting and agents, or you want the daily-habit machinery of streaks, ranks and duels to keep you coming back.
Choose both if you can justify two subscriptions. They are not really competing for the same hour.
If you want a read on where you stand before spending anything, the AI rank quiz is ten questions and about two minutes, with no signup and no card.
Iro AI turns ideas like the ones in this post into 5-minute exercises with feedback. Free tier, Pro from $4.17 a month ($49.99 a year, 7-day free trial).
Wondering is a learning app from Y Combinator's Summer 2026 batch that turns any topic into a personalised path of roughly three-minute lessons with visuals, generated podcasts and exercises. You can start from a topic, a link, a PDF or a description of your goal. It is in early access on iOS and web, built by a two-person team in San Francisco.
They are built for different jobs. Wondering is broad: any subject, personalised, strong at turning source material into a structured path. Iro AI is narrow and deep on using AI, and it grades your work rather than only explaining. If you want general curiosity satisfied, Wondering fits. If you want to be measurably better at prompting and AI tools, Iro fits.
Wondering is free to start, with Wondering Pro at $7.50 a month on annual billing, which works out to $90 a year. That figure is an early-supporter founding rate, so check what you are quoted. For comparison, Iro Pro is $49.99 a year or $9.99 a month, making Wondering roughly 80% more on annual billing.
Yes, and from the same kinds of input: a topic you type, a TikTok or YouTube link, a shared Instagram post, a screenshot or a PDF. Iro also presents lessons visually and has audio versions of lessons shipping shortly. The lasting difference is not the format but the depth of the AI curriculum and the Prompt Lab, which grades prompts you write against a real model.
No. At the time of writing it is in early access on iOS and the web only. Iro AI is on iOS and runs in any browser, with Android in development.
Iro AI, by design. It is a single-subject app with 29 paths, 477 lessons and 2,700+ exercises on AI tools, prompting, agents and automation, plus a Prompt Lab that grades prompts you write. A general learning app can explain AI concepts well, but it will not put you through graded reps on the skill.
Both are, in different ways. Wondering has personal streaks and Friend Streaks with up to three mutual friends, plus accuracy and activity tracking. Iro adds XP, six competitive ranks, live ELO-matched duels against other learners, daily challenges, chests and an evolving mascot. Wondering's system rewards consistency; Iro's rewards competition and visible progression.