TLDR; Duolingo Comprehension Feed turns short native-speaker videos into level-matched listening practice with tap-to-translate captions. Quick comprehension checks keep scrolling active, rewarding understanding with XP while automatically adding missed phrases to review.
I found Parrot on a YC startup rabbit hole, a three-person team pitching itself as "the TikTok for language learning," built entirely on the idea that scrolling short native-Spanish video clips teaches you faster than drills do.
One of the cofounders' bios reads, half-joking: "RIP Duolingo."
Spanish is the platform's flagship course: the largest and most invested course on the app, with roughly 236–239 units across 9 sections, one of only nine courses built out through B2 level as of April 2026[3], and dedicated Stories content. If Duolingo pilots anything experimental, Spanish is where it gets the deepest content investment first.
Monetization is subscription-led: Super Duolingo runs about $13.99/month or $95.99/year and removes ads, gives unlimited hearts, and unlocks personalized practice[3]. The free tier is ad-supported.
<aside> 📊
73% of self-study language app users say speaking exercises prepare them only "somewhat" or "very little" for real-world conversation, and 1 in 4 have never spoken their target language with a native or fluent speaker.[4]
Source: Preply, 2025 survey of self-learning language app users.
</aside>
Duolingo gets intermediate Spanish learners through vocabulary and grammar drills, but it doesn't have a format for understanding native speech at natural speed and scenarios (everything is animated!!)
Learners plateau, they can ace lessons and still freeze in a real conversation, and when they go looking for a fix, a growing set of video-based products (Parrot, Lingopie, FluentU) are positioned to catch them on the way out.
Independent learners describe the same wall from the outside.
One long-time user wrote a public post arguing that Duolingo's gamification optimizes for streak protection over real progress, and specifically floated watching a few minutes of target-language video as a better use of that time than another already-mastered review lesson[6].
And the market has already priced this pain in: Preply's 2025 survey found 54% of self-learning app users believe their apps limit their ability to progress specifically because there's no real-world engagement built in[4].
| Product | Format | Strength | Weakness |
|---|---|---|---|
| Duolingo (today) | Drills + AI conversation calls | 133M MAU, daily habit, gamified retention | No native-video comprehension surface at all |
| Parrot | Vertical short-form native video feed, Spanish only | Format is validated, users already scroll for hours | 3-person team, pre-seed, single language, unverified independent metrics, zero existing user base |
| Lingopie / FluentU | Full TV shows, movies, music videos | Deep content libraries, established for years | Long-form, appointment-viewing, not built for a scroll habit |
| Language Reactor | Netflix/YouTube overlay extension | Free, works on existing content | Manual setup, no curation, no learning path |
