frequently asked

what is madsio?

madsio is a language learning platform with five companions who befriend you first, then teach you. unlike apps that throw you into flashcards, our companions start in YOUR language, get to know your personality, and gradually introduce your target language through natural conversation. real voices. real pronunciation feedback. real relationship.

how is this different from duolingo or other language apps?

three ways. first: we start in your native language and build a friendship before teaching — no app does this. second: our companions actually LISTEN to your pronunciation with phoneme-level scoring, not fake string matching. third: every person gets a different experience because the companion adapts to your personality, anxiety level, and learning speed — not a one-size-fits-all curriculum.

what languages can i learn?

english (new york, london, sydney, general american, RP british, scottish, singapore), mandarin chinese (beijing standard, taiwanese), japanese (tokyo, kansai/osaka), korean (seoul, busan), spanish (mexico city, madrid, buenos aires, bogotá), french (parisian, québécois), german (hochdeutsch, austrian, swiss), and thai (bangkok central). we teach the specific accent you want, not just "the language."

are the companions real people?

they're characters with distinct personalities and teaching philosophies. kai is gentle and encouragement-first. atlas is direct and pushes hard. pen teaches real-world speech. sage connects language to culture. noor breaks everything into logical patterns. each has a different methodology — they genuinely teach differently, not just talk differently.

can i switch companions?

yes, anytime. your person profile carries across all companions — so if you start with kai and move to atlas, atlas already knows your level, your weak sounds, and your goals. we actually encourage switching: different companions unlock different aspects of your learning. on intensive plans, you can switch mid-session.

how does the companion know what level i'm at?

through conversation, not tests. during your first sessions, the companion naturally discovers your vocabulary range, grammar intuition, pronunciation patterns, and listening comprehension — just by talking with you. there's no placement test because real teachers don't start with tests. they start with "tell me about yourself."

does the companion actually speak out loud?

yes — with real, human-quality voice powered by ElevenLabs multilingual v2. each companion has their own unique voice with emotion, rhythm, and regional accent character. this is the same technology behind professional audiobooks and film dubbing. it does not sound robotic.

can the companion hear me speak?

yes — and this is the critical difference. when you speak, we capture your audio and run it through pronunciation assessment that scores each individual phoneme (sound). the companion can tell you "your 'th' sound scored 45% — listen to how i say it" and then demonstrate the correct pronunciation. this is real correction, not just transcription.

what's the deal with accents?

we believe "british english" is not specific enough. do you want london english? scottish english? RP? the accent you choose affects everything: vocabulary choices ("flat" vs "apartment"), pronunciation targets (dropped R's in british), and cultural references. your companion speaks in your chosen accent and coaches you toward it specifically.

why does the companion start in my language?

because research (Krashen's Affective Filter hypothesis, confirmed by 2025 neuro-ecological studies) shows that anxiety literally blocks language acquisition at a neurological level. by starting in your language, building trust, and understanding who you are, we lower that barrier. the target language gets introduced gradually, naturally, inside a conversation you're already enjoying.

what are the three phases?

phase 1 (befriend): sessions 1-3 are mostly in your native language. the companion gets to know you. phase 2 (bridge): sessions 4-10 blend both languages, shifting from 20% to 50% target language. phase 3 (immerse): session 11+, primarily in the target language with your native as a safety net. the transition is natural — you won't feel it happening.

do i need to be a complete beginner?

no — the system adapts. if you already have intermediate skills, the companion will detect this in the first session and adjust the phase progression. advanced learners might skip straight to phase 2 or 3. the phases aren't rigid steps — they're a framework that bends to match you.

what does the companion remember about me?

your companion builds a person profile through conversation: your goals, personality type, learning speed, weak pronunciation sounds, vocabulary knowledge, and session history. this data is stored securely and used only to personalize your learning. you can view, export, or delete your profile at any time in settings.

is my voice recorded?

voice audio is processed in real-time for transcription and pronunciation scoring, then discarded. we do not store voice recordings. only the text transcript and pronunciation scores are retained in your profile to track progress over time.