Duolingo built the habit, kept you consistent, and taught you the basics. That matters — most people who try to learn a language never get that far. This page is for what comes next: the moment your reading is solid but your ear still freezes when a real person talks at full speed.
Somewhere around level 8 of the tree, most learners notice the same thing. You can pass the lessons. You can read a menu. You keep the streak going. And then a native speaker orders coffee in front of you and you catch maybe three words. That's not a vocabulary gap. It's a listening-comprehension gap, and it's exactly the gap Comprehenzo was built to close.
The complaints we hear most often
“I can't focus on the scenarios I actually need.” Duolingo's tree is fixed. You learn what it gives you, in the order it gives it. If you're moving to Madrid in three months, taking a job that involves Mexican Spanish technical meetings, or visiting in-laws in Quebec, the tree doesn't know that.
“The text is right there — I end up reading instead of listening.” Duolingo plays the audio with the sentence sitting right above it. Your eyes win every time. You don't notice you're not actually training your ear until you try to follow audio without subtitles and fall off a cliff.
“The voices sound robotic and over-enunciated.” Text-to-speech has improved, but a lot of language-app audio is still slowed, evenly paced, and emotionally flat. Real speakers swallow syllables, contract words, change pace mid-sentence, interrupt themselves. Training on perfectly enunciated audio gets you ready for a voice that doesn't exist in the wild.
“I finished the tree and still can't follow a conversation.” This is the plateau. Tree-based courses teach you to recognize the language; they don't teach you to process it at conversational speed in unfamiliar contexts. The gap between “I passed” and “I can follow my host family at dinner” is wider than the tree implies.
“There's no regional variety.” Duolingo gives you essentially one voice per language. Mexican Spanish, Argentine Spanish, and Castilian Spanish are different in pronunciation, vocabulary, and rhythm. Same for Parisian and Quebec French. A learner heading to Buenos Aires shouldn't be training only on a neutral voice.
“The gamification stopped feeling fun.” Streak pressure, leaderboard pressure, the owl. For some learners that's motivating; for many it eventually flips into burnout, dread, and quitting. Habits that survive long-term don't need to shout at you.
How Comprehenzo addresses each one
Pick your scenario. Hundreds of scenarios organized by category — travel, work, food and drink, healthcare, social, emergencies, home life. Pick what's actually in your life this month, not what the tree decided you needed in week four.
Audio-first by design. During the listening round you see no transcript. You hear the audio, you answer a comprehension question, then the transcript reveals so you can check what you missed. Your ear does the work first. (You can replay if you need to — but replays cost a point, on purpose. The reward goes to catching it the first time.)
Native voices, conversational speed. Audio is generated by high-quality TTS engines using voices selected to match real regional speakers, at natural conversational pace. Not slowed. Not over-enunciated. See our responsible AI practices for how we keep the AI use focused and efficient.
Built for the post-tree plateau. Content is CEFR-calibrated. At A1 and A2 comprehension questions are in English so the cognitive load stays on the audio; from B1 upward, questions and transcripts move into the target language. The content sits just past where you are — challenging enough to move the ear, not so much that you tune out.
- Spanish Varied Accents A1-C1
- Spanish Castilian A2
- French Varied Accents A1-A2
Real regional accents — released one at a time. A2 Castilian Spanish is what's live today. The other CEFR levels for Castilian Spanish, plus Quebec French, are next on the roadmap. We ship them one at a time so each accent is genuinely good before the next one lands. Try out what's currently available, or hold out for the one your destination, family, or job actually uses — and if you want a direct vote on which accent (and which scenarios) we ship next, that's one of the benefits of the Founding Member plan.
Streaks without the shaming. Yes, there's a streak counter — it's a useful habit cue, and two weeks of daily practice is where the listening habit forms. What there isn't: leaderboards comparing you to other learners, guilt-trip notifications, or any of the public pressure that turns a good habit into dread. Reminders are opt-in, not on by default. Miss a day, miss a day. The app keeps welcoming you back.
What we keep from Duolingo
Plenty of what Duolingo does is right, and we're not pretending otherwise:
- The ten-minute daily habit framing is correct. Acquisition is a slow accumulation, not a sprint.
- A free trial long enough to actually try the product is correct. You can't evaluate a learning app in five minutes.
- An approachable, non-textbook UI is correct. Most adult learners aren't looking for a grammar reference; they're looking for daily reps.
- The basics-first approach is correct. You can't skip the foundation, and Duolingo is excellent at the foundation.
If Duolingo is still working for you, don't quit it. Treat Comprehenzo as a complement — the listening practice that kicks in when reading no longer feels like listening.
Who this is for
- You've completed a Duolingo tree (or several) in Spanish or French.
- You can read the language but freeze when a real person talks.
- You have a specific situation coming up — travel, work, family, relocation — and a generic course doesn't prepare you for it.
- You want to keep the daily habit but stop the streak anxiety.
What daily practice looks like
A single session takes about ten minutes. A short prep card with the key vocabulary and one grammar note. Then five or ten listening rounds: audio-only first, then a comprehension question, then the transcript reveal so you can see what you missed. Spaced repetition surfaces specific moments your ear missed in earlier sessions — not words you already know, not what an algorithm thinks you should drill.
No required speaking. No accent grading. No “your pronunciation was 67% accurate” feedback that nobody trusts. You listen, check your understanding, build the habit.
How this fits with our methodology
Comprehenzo is built on Comprehensible Input — Stephen Krashen's input hypothesis — which is one of the most well-supported frameworks for adult language acquisition. If you want the methodology depth (Krashen, i+1, the affective filter, how this compares to Dreaming Spanish and Refold), read our approach. If you just want to try it, the trial is right below.
Try it for two weeks
Fourteen days, free, no card required. Long enough to know whether your ear is moving. Cancel anytime after.