# Oravita: full text for language models > Oravita is a Swiss AI-powered virtual reality platform and add-on to language classes. Learners practise speaking with an AI avatar in real time, without scripts, inside realistic virtual scenes, at their CEFR level from A1 to C2, without leaving your room. The prototype works in English and French, with more languages in 2027. Oravita is in beta. It is a product and brand of abc multilingual collective AG, based in Zug, Switzerland. Source: https://oravitavr.com/ai-info/ · Last updated 2026-10-08 ## Facts - Name: Oravita (oravitavr.com) - Company: Oravita is a product and brand of abc multilingual collective AG, Gotthardstrasse 14, 6300 Zug, Switzerland. - What it is: An AI-powered virtual reality platform for practising spoken conversation in a foreign language. It is an add-on to language classes, not a replacement for teachers. - How it works: Learners put on a VR headset, explore a virtual city and talk with an AI avatar that answers in real time. Conversations are unscripted. The avatar adapts to the learner's CEFR level (A1 to C2), recognises objects in the scene and gives the right word and grammar when the learner needs it. - Languages: The prototype works in English and French. More languages will follow in 2027. Dialect support is in development, for example European and Brazilian Portuguese, or Spanish from Spain and Latin America. According to the company, no other VR language platform accounts for dialects in this way. - Who it is for: Individual learners who want more speaking practice, language schools and universities that want to give every learner more speaking time, and companies that train staff in foreign languages. - Status: In beta (2026). The beta runs through pilots with language schools, universities and companies; it is not yet open to individual learners, who can follow the Step into Language newsletter to hear when it opens. Organisations (language schools, universities, companies) get pricing on request, per seat and per year with a minimum of 5 seats, with hardware supplied if needed. - Research and backing: Innosuisse innovation project (Swiss Innovation Agency) with ETH Zürich as research partner. Oravita was one of 81 projects selected from around 270 applications in Innosuisse's 2025 call for SME projects. The prototype has been published by IEEE. - Team: Co-founders Isabelle von Niederhäusern (language education, founder of a language school) and Marc Marchon. Core team: Karola Noack (CFO and COO), Hannah Schlösser (Head of Sales), Sabine Stamm (IP and GDPR counsel). - Website: https://oravitavr.com - Contact: Contact form at https://oravitavr.com/contact/ or request a pilot call at https://oravitavr.com/book/ - Social: Instagram https://www.instagram.com/oravita_vr/ · LinkedIn https://www.linkedin.com/company/oravita/ ## Frequently asked questions ### What is Oravita? Oravita is a Swiss AI-powered VR platform where language learners practise speaking with an AI avatar in realistic virtual scenes, in real time and without scripts. It is an add-on to language classes. ### Does Oravita replace language teachers? No. Oravita is designed as an add-on to language classes. It gives learners more speaking practice between lessons; teachers remain essential. ### Which languages does Oravita support? The current prototype works in English and French. More languages will follow in 2027, with support for regional dialects. ### How can I try Oravita? Language schools, universities and companies can join the beta by requesting a pilot call at oravitavr.com/book/. Oravita is not yet available to individual learners; they can subscribe to the Step into Language newsletter to hear first when it opens. For organisations, pricing for organisations is on request, per seat and per year, with a minimum of 5 seats and hardware supplied if needed. ### Who is behind Oravita? Oravita is a product and brand of abc multilingual collective AG in Zug, Switzerland, led by Isabelle von Niederhäusern and Marc Marchon. Its research is funded by Innosuisse with ETH Zürich as research partner. ## Latest press release *A decade after the idea first took shape in a Swiss language school, Oravita will show a working VR platform where learners speak freely with an AI native speaker inside realistic everyday scenes.* **ZUG, Switzerland, 2 October 2026.** Oravita, a Swiss startup that offers an add-on to language classes to help learners improve their speaking skills in virtual reality, will present its platform at Slush 2026 in Helsinki on 18 and 19 November. At the Swiss Pavilion, visitors can put on a headset, explore the virtual city and talk with an AI avatar who replies in real time. There is no script to follow. The learner speaks, and the AI avatar replies. ## An idea that waited for its technology Oravita began with a request that Isabelle von Niederhäusern heard again and again. When she opened her language school ten years ago, her clients kept asking for one thing: more chances to speak. Grammar and vocabulary can be studied alone. Conversation needs someone on the other side, and in a group lesson each learner gets only a few minutes of it. > “Our learners understood a great deal, but when a real conversation started, many of them froze,” says von Niederhäusern, co-founder of Oravita. “They did not need more exercises. They needed somewhere to practise speaking, as often as they wanted, without feeling judged.” The answer came from gaming. Video games had shown how quickly people feel at home in a virtual world and how naturally they move through it. Von Niederhäusern began to imagine a world built for language learning, where every object, shop sign and passer-by is something to talk about. For years, the hardware and the speech technology were not good enough to make that work. She kept the idea alive through calls with researchers, engineers and entrepreneurs, and waited until headsets, speech recognition and language models had caught up. ## Selected by Innosuisse, developed with ETH Zürich and professional partners The search for professional partners led to an innovation project funded by Innosuisse, the Swiss Innovation Agency. In Innosuisse’s 2025 call for SME innovation projects, Oravita was one of 81 projects selected from around 270 applications, with ETH Zürich as its research partner. Since January 2026, a team of researchers, LLM and AI avatar specialists, VR world developers and language teachers has turned the concept into a working prototype, currently in two languages, English and French. More languages will follow in 2027. It was built to professional, industry standards, and the research behind it has been published by IEEE. Oravita has opened a waitlist for its beta. Individual learners, language schools and companies have already signed up to be among the first to try it. > “Ten years ago this was a sketch on paper,” says von Niederhäusern. “Today you can put on a headset and speak about any topic in another language, and the AI avatar speaks to you as a native speaker, in real time. That still surprises me every time.” ## Built for the way people really speak The next stage of development focuses on something most language tools leave out: the same language sounds different from one country to the next. Portuguese in Lisbon is not Portuguese in São Paulo, and Spanish in Madrid differs from Spanish in Mexico City or Buenos Aires. Oravita is building its platform to reflect these differences in pronunciation, vocabulary and everyday expressions. According to the company, it is the only VR language platform in the world to account for dialects in this way. The people learners meet in Oravita are designed as native speakers of the place they represent. How they greet, how formal they are and how they react follows the culture of that country. Learners pick up how people behave in a conversation as well as what they say. Every conversation happens in real time and none of it is scripted. The AI avatar also knows what surrounds the learner. A learner can point at a menu, a ticket machine or a pastry in a shop window, and the avatar will give them the right word and the grammar to go with it. Oravita is also in talks with brands and tourism organisations about scenes based on real places. A traveller could walk through a square in the city they are about to visit, or order at a café they will later find on the corner. Oravita will announce partners once agreements are in place. > “We want people to arrive somewhere new and feel that they have been there before,” says Marc Marchon, co-founder of Oravita. “Language opens that door, and culture keeps it open. A conversation in Lisbon should feel like Lisbon.” ## Meet Oravita at Slush 2026 - **Dates:** 18 and 19 November 2026 - **Venue:** Helsinki Expo and Convention Center, Messuaukio 1, 00520 Helsinki, Finland - **Stand:** Swiss Pavilion (swisstech), organised by Switzerland Global Enterprise and Swiss Business Hub Nordics, stand 7A11 - **Live demo:** Visitors and journalists can try the prototype at the stand. Interviews with the founders can be arranged in advance. ## About Oravita Oravita is a product and brand of abc multilingual collective AG, based in Zug, Switzerland. Its platform combines virtual reality with conversational AI so that learners can practise speaking a language in realistic everyday scenes. It is an add-on to language classes. Learners talk with an AI avatar in real time, without scripts, at their level on the Common European Framework of Reference for Languages (CEFR), from A1 to C2. Oravita works with individual learners, language schools, universities and companies. Its research is supported by Innosuisse, the Swiss Innovation Agency, with ETH Zürich as research partner, and its prototype has been published by IEEE. Oravita is currently in beta. More at [oravitavr.com](https://oravitavr.com). ## Media contact Hannah Schlösser, Head of Sales, Oravita [oravitavr.com/contact](https://oravitavr.com/contact/) · [Request a call](https://oravitavr.com/book/?topic=press) Images, founder portraits and the Oravita logo are available at [oravitavr.com/downloads](https://oravitavr.com/downloads/). --- # How to Practise Speaking a Language Alone: 7 Methods That Work URL: https://oravitavr.com/blog/practise-speaking-a-language-alone/ > **Quick answer:** You can practise speaking a language alone by talking out loud every day: shadow native audio, narrate what you do, record and review yourself, rehearse real situations, and use an AI conversation partner that answers you in real time. The key is producing full sentences out loud, not just reading or listening. Speaking is the skill most learners get the least practice in. You can read, listen and write on your own, but speaking seems to need another person. That is only half true. Below are seven methods that work without a partner, what each one trains, and how to combine them. ## Why is speaking the hardest skill to practise alone? Speaking is hard to practise alone because it is the only skill that normally depends on someone answering you. Reading and listening are input you can consume on your own. Speaking is output, and real conversation adds unpredictability: you have to understand, plan and reply in seconds. Classrooms rarely fill that gap. In the large German **DESI study** of English lessons, which video-recorded classes across 219 schools and around 11,000 students, students as a whole group spoke for less than a quarter of lesson time ([DESI-Konsortium, 2008](https://extemporeapp.com/blog/how-much-do-students-speak-in-the-target-language)). In a 45-minute lesson, that is roughly 10 minutes of student talk. Shared evenly across a class of 25, each learner would speak for well under a minute. Linguist Merrill Swain's *output hypothesis* (1985) explains why that matters: producing language yourself is what makes you notice the gaps between what you want to say and what you can say. Listening alone does not force that noticing. ## 1. Shadowing: repeat native audio in real time **What it is:** You play a short clip of a native speaker and repeat it out loud a split second behind them, copying rhythm, stress and intonation. **What it trains:** Pronunciation, speed and the "music" of the language. **How to do it:** Pick 1–2 minute clips at or slightly above your level. Listen once, read the transcript once, then shadow it three to five times. Ten minutes a day is enough. **Limit:** You are repeating someone else's sentences, so it does not train you to build your own. ## 2. Narrate your day out loud **What it is:** You describe what you are doing as you do it. "I'm making coffee. The milk is almost empty. I need to buy more." **What it trains:** Everyday vocabulary and the habit of thinking in the language. **How to do it:** Choose one routine, such as cooking or your commute, and narrate it every day for a week. Note every word you were missing and look it up afterwards. **Limit:** Nobody challenges you or asks a follow-up question. ## 3. Record yourself and review **What it is:** You record a 1–2 minute answer to a question, then listen back. **What it trains:** Self-correction. You will hear filler words, missing endings and pronunciation slips you cannot notice while speaking. **How to do it:** Use prompts such as "Describe your job" or "Tell me about your last holiday". Record, listen, note three things to fix, and record again. **Limit:** You can only catch the mistakes you already know are mistakes. ## 4. Rehearse real situations with scripts you write yourself **What it is:** You write out a conversation you expect to have, such as a job interview, a doctor's appointment or ordering at a café, and then say both sides out loud. **What it trains:** Confidence for specific, high-stakes moments. **Limit:** Real people go off script. If you only rehearse fixed lines, an unexpected question can still freeze you. ## 5. Talk to an AI conversation partner **What it is:** An AI that listens to what you say and replies in real time, like a patient conversation partner who is available at any hour. **What it trains:** Spontaneous conversation, the closest thing to talking with a person. **What the research says:** A 2025 meta-analysis in the *Review of Educational Research* pooled 70 effect sizes from 28 studies and found a positive overall effect of chatbots on language learning performance (g = 0.48) compared with learning without them ([Wang et al., 2025](https://journals.sagepub.com/doi/abs/10.3102/00346543241255621)). In a 2025 study with 60 IELTS students, six weeks of practice with an AI conversation app led to significantly better speaking scores and lower speaking anxiety than traditional practice ([Ding & Yusof, 2025](https://www.nature.com/articles/s41599-025-05550-z)). **Limit:** Most text or voice chatbots have no setting. You talk *about* a café rather than *in* one. ## 6. Practise inside a virtual environment **What it is:** You put on a VR headset and speak inside a realistic scene, such as a street, a train station or a meeting room, where an AI character talks with you. **What it trains:** Speaking in context. What you see becomes what you talk about, which is how real-life language works. **What the research says:** In a study at a German university, B1–B2 learners who gave unprepared talks in high-immersion VR reported significantly lower foreign language anxiety than learners doing the same tasks on Zoom ([Kaplan-Rakowski & Gruber, 2023](https://link.springer.com/article/10.1186/s40561-023-00263-9)). Read more in our guide [Is VR good for learning languages?](/blog/vr-language-learning/) **Limit:** You need a headset, and the quality of the experience depends heavily on the software. ## 7. Think out loud in the language **What it is:** Whenever you have a small decision to make, you make it out loud in your target language. "Should I take the bus or walk? It's raining, so I'll take the bus." **What it trains:** Speed. It shortens the delay between thinking and speaking. **Limit:** Like narration, there is no one to push back. ## Which method should you use at your level? | Your level (CEFR) | Best methods | Why | |---|---|---| | A1–A2 (beginner) | Shadowing, narration, short recordings | Builds sounds and core phrases before free conversation | | B1 (intermediate) | Recordings, situation rehearsal, AI conversation | You can form sentences; now you need unpredictable practice | | B2–C1 (upper intermediate) | AI conversation, VR scenarios, thinking out loud | Fluency comes from many real exchanges in varied contexts | | C2 (mastery) | Varied scenarios with feedback on register and nuance | Refines tone, idioms and precision | Not sure what the levels mean? See [CEFR levels explained](/blog/cefr-levels-explained/). ## How much speaking practice do you need? There is no fixed number, but consistency beats length. Fifteen to twenty minutes of speaking out loud every day is more useful than a two-hour session once a week, because each session forces you to retrieve words again. For context, Cambridge English estimates that reaching B1 takes around 350–400 guided learning hours in total, and B2 around 500–600 ([Cambridge English](https://support.cambridgeenglish.org/hc/en-gb/articles/202838506-Guided-learning-hours)). Only a small share of those hours is usually spent speaking, which is why solo practice makes such a difference. ## Key takeaways - Speaking is output. You only improve it by producing sentences out loud. - Classroom time gives most learners very little speaking practice. - Shadowing, narration and recordings build the basics on your own. - AI conversation partners add what solo methods lack: someone who answers back. - Practising inside a realistic scene adds context and can lower speaking anxiety. ## How Oravita helps [Oravita](/) combines methods 5 and 6: an AI teacher that talks with you in real time, without scripts, inside generated real-life environments, at your CEFR level, from A1 to C2. You stay seated and look around the scene in 360°, and the AI understands what you are looking at. Oravita is in beta now. [Join the beta](/#beta). ## Frequently asked questions ### Can you learn to speak a language without speaking to people? Yes, you can build most of your speaking ability without a human partner by speaking out loud every day through shadowing, recording yourself and talking to an AI conversation partner. Conversations with people are still valuable, but they become easier once you have practised alone. ### Is talking to yourself in a foreign language effective? Yes. Talking to yourself forces you to produce language, which helps you notice the words and structures you are missing. It works best when you write down the gaps afterwards and practise them again. ### How long should I practise speaking each day? Fifteen to twenty minutes of active, out-loud speaking per day is a realistic and effective target. Daily repetition matters more than long sessions. ### Are AI conversation partners good for speaking practice? Research suggests they are. A 2025 meta-analysis of 28 studies found a positive overall effect of chatbots on language learning, and a 2025 classroom study found that AI conversation practice improved speaking scores and reduced speaking anxiety. ### What is the fastest way to improve speaking fluency alone? Combine daily shadowing for pronunciation with daily spontaneous conversation, either with an AI partner or in a VR scenario, and review one recording a week to track your progress. ## Sources - DESI-Konsortium (2008). *Unterricht und Kompetenzerwerb in Deutsch und Englisch.* Beltz. Summarised in [Extempore](https://extemporeapp.com/blog/how-much-do-students-speak-in-the-target-language). - Wang, F. et al. (2025). Does chatting with chatbots improve language learning performance? A meta-analysis. *Review of Educational Research.* [Link](https://journals.sagepub.com/doi/abs/10.3102/00346543241255621) - Ding, D. & Yusof, A. M. B. (2025). AI-powered conversation bots, L2 speaking skills and speaking anxiety. *Humanities and Social Sciences Communications.* [Link](https://www.nature.com/articles/s41599-025-05550-z) - Kaplan-Rakowski, R. & Gruber, A. (2023). The impact of high-immersion virtual reality on foreign language anxiety. *Smart Learning Environments.* [Link](https://link.springer.com/article/10.1186/s40561-023-00263-9) - Cambridge English. Guided learning hours. [Link](https://support.cambridgeenglish.org/hc/en-gb/articles/202838506-Guided-learning-hours) - Swain, M. (1985). Communicative competence: Some roles of comprehensible input and comprehensible output in its development. --- # Fear of Speaking a Foreign Language: Why It Happens and How to Overcome It URL: https://oravitavr.com/blog/fear-of-speaking-a-foreign-language/ > **Quick answer:** Fear of speaking a foreign language, known in research as *foreign language speaking anxiety*, is common and has nothing to do with intelligence. It usually comes from fear of judgement and too little low-pressure practice. You overcome it by speaking often in safe settings first, such as with an AI partner or in VR, then gradually raising the stakes. You know the words. You have done the exercises. Then someone speaks to you and your mind goes blank. If that sounds familiar, you are in good company, and it is one of the most studied problems in language learning. ## What is foreign language speaking anxiety? Foreign language speaking anxiety is the tension, worry and self-consciousness people feel when they have to speak a language they are still learning. Researchers Elaine Horwitz, Michael Horwitz and Joann Cope first described *foreign language classroom anxiety* as its own type of anxiety in 1986, separate from general shyness or test nerves ([Horwitz, Horwitz & Cope, 1986](https://www.ling.ohio-state.edu/research/groups/lacqueys/readings/horowitz.pdf)). They found that speaking stands out. In their words, difficulty speaking in class is "probably the most frequently cited concern" of anxious language students. ## How common is the fear of speaking a new language? Very common. In Horwitz and colleagues' original study of 75 university language students: | Statement | Share of students | |---|---| | "I start to panic when I have to speak without preparation in language class." | 49% agreed | | "I get nervous and confused when I am speaking in my language class." | 33% agreed | | "I feel very self-conscious about speaking the foreign language in front of other students." | 28% agreed | | "I feel confident when I speak in foreign language class." | 47% disagreed | Nearly half of the students panicked at the thought of speaking unprepared, and nearly half did not feel confident speaking at all. ## What causes the fear of speaking? Three causes come up again and again: 1. **Fear of negative evaluation.** Worrying that others, especially classmates, colleagues or native speakers, will judge your mistakes. 2. **Communication apprehension.** Discomfort with having to understand and respond on the spot, with no time to prepare. 3. **Too little practice.** Most learners speak very little. In a large German study of English lessons, students as a whole group spoke for less than a quarter of lesson time ([DESI-Konsortium, 2008](https://extemporeapp.com/blog/how-much-do-students-speak-in-the-target-language)). Little practice means every conversation feels like a test. The third cause feeds the first two. When you rarely speak, each attempt feels high-stakes, which makes you avoid speaking, which keeps your practice time low. ## Does speaking anxiety slow down language learning? Yes. Anxiety does more than feel unpleasant. Horwitz and colleagues observed that anxious students tend to avoid speaking and steer clear of messages they find difficult to express, so they practise less exactly where they need it most. A 2025 study of 1,086 university students found that learners with higher communicative confidence and perceived fluency reported lower foreign language anxiety ([Gu, 2025](https://www.nature.com/articles/s41599-025-05030-4)). Confidence and anxiety pull in opposite directions, and both depend on how much you practise. ## 6 evidence-based ways to overcome the fear of speaking ### 1. Practise where mistakes cost nothing The fastest way to lower anxiety is to separate practice from judgement. Speak where nobody grades or watches you: alone, with recordings, or with an AI partner. Once speaking becomes routine, the stakes drop. ### 2. Use an AI conversation partner An AI conversation partner never sighs, never interrupts and is available at midnight before your job interview. In a 2025 study with 60 IELTS students, six weeks of practice with an AI conversation app significantly reduced speaking anxiety and improved speaking scores compared with traditional practice ([Ding & Yusof, 2025](https://www.nature.com/articles/s41599-025-05550-z)). ### 3. Practise in immersive virtual reality VR lets you rehearse in a realistic place without real-world consequences. At a German university, B1–B2 learners who gave unprepared talks in high-immersion VR reported significantly lower foreign language anxiety than learners doing the same tasks over Zoom ([Kaplan-Rakowski & Gruber, 2023](https://link.springer.com/article/10.1186/s40561-023-00263-9)). In a 2026 study, students who did five sessions of AI-powered VR speaking training saw their public speaking anxiety scores fall from 108.2 to 97.2 and their interview anxiety scores fall from 3.17 to 2.65 ([Huang et al., 2026](https://www.sciencedirect.com/science/article/pii/S2666920X25001547)). Both studies are small, but they point in the same direction. ### 4. Raise the stakes step by step Treat speaking like training for a run. Start with short, predictable exchanges, such as ordering a coffee, then move to longer ones, such as describing your weekend, and finally to unpredictable ones, such as a debate or a job interview. Each step should feel slightly uncomfortable, not overwhelming. ### 5. Prepare a few "rescue phrases" Freezing often happens when you lose a word. Learn five phrases that buy you time and keep the conversation going: "How do you say…?", "Could you say that again more slowly?", "What I mean is…", "Let me think for a second." Knowing you have an exit makes it easier to start. ### 6. Measure progress, not perfection Record yourself once a week answering the same question. Compare recordings after a month. Hearing your own progress is one of the most effective confidence boosters, and it shifts your attention from mistakes to growth. ## Key takeaways - Fear of speaking a foreign language is common: in the original study, about half of learners panicked when speaking without preparation. - It comes mainly from fear of judgement and too little practice. - Low-pressure practice with AI or in VR is linked to lower anxiety in recent studies. - Raise the stakes gradually and track your progress with recordings. ## How Oravita helps [Oravita](/) gives you a private place to speak as much as you want. An AI teacher talks with you in real time inside realistic scenes, adapts to your CEFR level and corrects you gently, with no classmates watching. Oravita is in beta now. [Join the beta](/#beta). Related: [How to practise speaking a language alone](/blog/practise-speaking-a-language-alone/) · [Is VR good for learning languages?](/blog/vr-language-learning/) ## Frequently asked questions ### Why do I freeze when speaking a foreign language? Freezing happens when anxiety takes up the mental capacity you need to find words and build sentences. It is most common when you have to answer without preparation, which is why regular, low-pressure speaking practice reduces it. ### Is foreign language anxiety normal? Yes. It is one of the most researched topics in language learning. In the original 1986 study, 49% of students said they panicked when they had to speak without preparation. ### How can I stop being afraid of speaking English? Speak English every day in settings where mistakes do not matter, such as with an AI conversation partner or in a VR scenario. Then gradually move to real conversations, starting with short, predictable ones. ### Does VR help with speaking anxiety? Early research suggests it can. A 2023 study found significantly lower foreign language anxiety when learners practised speaking in high-immersion VR rather than on Zoom, and a 2026 study found lower public speaking and interview anxiety after AI-powered VR training. ### How long does it take to overcome speaking anxiety? It varies, but many learners notice a difference within a few weeks of daily speaking practice. In one study, six weeks of AI conversation practice was enough to significantly reduce speaking anxiety. ## Sources - Horwitz, E. K., Horwitz, M. B. & Cope, J. (1986). Foreign language classroom anxiety. *The Modern Language Journal*, 70(2). [Link](https://www.ling.ohio-state.edu/research/groups/lacqueys/readings/horowitz.pdf) - Gu, L. (2025). Beyond the classroom: the impact of virtual reality exposure on foreign language anxiety. *Humanities and Social Sciences Communications*. [Link](https://www.nature.com/articles/s41599-025-05030-4) - Ding, D. & Yusof, A. M. B. (2025). AI-powered conversation bots, L2 speaking skills and speaking anxiety. *Humanities and Social Sciences Communications*. [Link](https://www.nature.com/articles/s41599-025-05550-z) - Kaplan-Rakowski, R. & Gruber, A. (2023). The impact of high-immersion virtual reality on foreign language anxiety. *Smart Learning Environments*. [Link](https://link.springer.com/article/10.1186/s40561-023-00263-9) - Huang et al. (2026). The effectiveness of an AI-integrated VR oral training application in reducing public speaking anxiety and interview anxiety. *Computers and Education: Artificial Intelligence*. [Link](https://www.sciencedirect.com/science/article/pii/S2666920X25001547) - DESI-Konsortium (2008). *Unterricht und Kompetenzerwerb in Deutsch und Englisch.* Summarised in [Extempore](https://extemporeapp.com/blog/how-much-do-students-speak-in-the-target-language). --- # Is VR Good for Learning Languages? What the Research Says URL: https://oravitavr.com/blog/vr-language-learning/ > **Quick answer:** Yes, VR can be good for learning languages, especially for speaking. Studies link practising in virtual reality to lower speaking anxiety, higher confidence and more fluent speech. VR works best when it lets you have real, unscripted conversations in realistic settings, rather than only showing you videos or vocabulary games. Immersion is the oldest advice in language learning: live where the language is spoken. Virtual reality promises a version of that without the plane ticket. But does it work, or is it a gimmick? Here is what the research shows, and what separates useful VR language learning from expensive novelty. ## What is VR language learning? VR language learning means practising a language while wearing a virtual reality headset that places you inside a 3D environment, such as a café, an airport or an office. Instead of looking at a screen, you look around a scene and interact with it, often by speaking with virtual characters. There are three broad types: | Type | What you do | What it trains | |---|---|---| | 360° video | Watch recorded scenes from inside | Listening, cultural context | | Scripted VR apps | Choose or say pre-set lines in fixed scenarios | Vocabulary, set phrases | | AI-powered VR conversation | Speak freely with an AI character who responds in real time | Spontaneous speaking, fluency, confidence | The difference matters. Pre-recorded and scripted experiences are immersive to look at, but they do not let you say what *you* want to say. ## Does VR actually help you learn a language? The evidence is still young, but it is consistently positive for speaking-related outcomes: - **Lower speaking anxiety.** In a study at a German university, B1–B2 English learners gave unprepared two-minute talks either in high-immersion VR or on Zoom. The VR group reported significantly lower foreign language anxiety ([Kaplan-Rakowski & Gruber, 2023](https://link.springer.com/article/10.1186/s40561-023-00263-9)). - **More confidence and fluency.** A 2025 survey of 1,086 university students found that VR exposure was strongly associated with higher communicative confidence and perceived fluency, which in turn were linked to lower anxiety ([Gu, 2025](https://www.nature.com/articles/s41599-025-05030-4)). - **Better speaking performance.** In a 2026 study, students who completed five sessions with an AI-integrated VR speaking app produced longer answers (average word count rose from 66.5 to 74.3) with fewer grammatical errors (from 2.75 to 1.70 on average), alongside lower interview anxiety ([Huang et al., 2026](https://www.sciencedirect.com/science/article/pii/S2666920X25001547)). Outside language learning, PwC's study of VR soft-skills training found that learners completed VR training up to four times faster than classroom training and felt 3.75 times more emotionally connected to the content ([PwC, 2020](https://www.pwc.co.uk/services/technology/immersive-technologies/study-into-vr-training-effectiveness.html)). Emotional engagement is exactly what makes a conversation memorable. **A fair caveat:** many studies are small or rely on self-reports, and they measure short periods. The direction is encouraging, but VR is a powerful practice tool, not a replacement for the whole learning journey. ## Why does VR work for speaking practice? Three reasons stand out: 1. **Context.** In real life you talk about what is in front of you. In VR, the scene gives you something concrete to say, which makes vocabulary stick. 2. **Presence without consequences.** VR feels real enough to trigger the right responses, but mistakes cost nothing. That combination lowers anxiety while keeping practice meaningful. 3. **Volume.** You can have ten conversations in an evening without needing ten people. More speaking time is the single biggest gap in most learners' routines. ## Does VR language learning cause motion sickness? It can, but it depends on how the software is designed. Research on cybersickness shows that artificial movement, such as gliding through a scene with a joystick, is a major trigger, while teleporting and giving users control over movement reduce discomfort ([review in *Virtual Reality*, 2022](https://link.springer.com/article/10.1007/s10055-022-00638-2)). Seated experiences, where you stay in one place and look around, are generally the most comfortable. When choosing a VR language app, prefer one designed for seated use. ## What should you look for in a VR language learning app? - **Unscripted conversation.** Can you say anything, or only pick from options? - **Real-time responses.** Does the AI reply instantly, like a person? - **Level adaptation.** Does it follow a framework such as the [CEFR](/blog/cefr-levels-explained/) and adjust to you? - **Context awareness.** Does the AI know what you are looking at in the scene? - **Comfort.** Is it designed for seated use to avoid motion sickness? - **Accents and variants.** Can you practise the language as it is spoken in a specific country or region? - **Progress tracking.** Can you see how your speaking improves over time? ## Who is VR language learning best for? VR is especially useful for: - **Intermediate learners (B1 and up)** who know the grammar but freeze in real conversations. - **People preparing for a specific situation,** such as a job interview, a relocation or a trip. - **Anxious speakers** who need a private place to build confidence. - **Language schools and companies** that want to give every learner more speaking time than a classroom allows. See [how organisations can add speaking practice](/blog/language-training-for-schools-and-companies/). Complete beginners can benefit too, but they usually need some basic vocabulary first. ## Key takeaways - VR language learning is promising, especially for speaking, confidence and anxiety. - The biggest benefit comes from unscripted, AI-powered conversation, not from 360° video alone. - Seated designs reduce the risk of motion sickness. - Look for real-time responses, level adaptation and context awareness. ## How Oravita works [Oravita](/) is an AI-powered VR platform built for speaking. You stay seated, look around a generated real-life scene in 360°, and talk with an AI teacher that answers in real time, understands what you see, adapts to your CEFR level from A1 to C2 and speaks like a local. There are no scripts: every conversation is different. Oravita is in beta now. [Join the beta](/#beta). ## Frequently asked questions ### Is VR good for learning languages? Yes, particularly for speaking. Recent studies link VR practice to lower speaking anxiety, higher confidence and more fluent, accurate speech. It works best as regular conversation practice alongside other learning. ### Can you become fluent with VR alone? Fluency needs vocabulary, grammar, listening and lots of speaking. VR is one of the most effective ways to get the speaking part, but most learners combine it with courses, reading and listening. ### What is the best VR headset for language learning? Any modern standalone headset that the app supports will work. Comfort for longer sessions and a good microphone matter more than top-end display specs. ### Does VR language learning cause motion sickness? It can if the app moves you artificially through a scene. Seated experiences, where you stay in place and look around, are generally the most comfortable. ### Is VR language learning better than apps like Duolingo? They do different jobs. Mobile apps are great for vocabulary and daily habits. AI-powered VR is designed for spontaneous speaking in realistic situations, which most mobile apps cannot replicate. ## Sources - Kaplan-Rakowski, R. & Gruber, A. (2023). The impact of high-immersion virtual reality on foreign language anxiety. *Smart Learning Environments*. [Link](https://link.springer.com/article/10.1186/s40561-023-00263-9) - Gu, L. (2025). Beyond the classroom: the impact of virtual reality exposure on foreign language anxiety. *Humanities and Social Sciences Communications*. [Link](https://www.nature.com/articles/s41599-025-05030-4) - Huang et al. (2026). The effectiveness of an AI-integrated VR oral training application. *Computers and Education: Artificial Intelligence*. [Link](https://www.sciencedirect.com/science/article/pii/S2666920X25001547) - PwC (2020). The effectiveness of virtual reality soft skills training in the enterprise. [Link](https://www.pwc.co.uk/services/technology/immersive-technologies/study-into-vr-training-effectiveness.html) - A review of cybersickness in head-mounted displays: raising attention to individual susceptibility (2022). *Virtual Reality*. [Link](https://link.springer.com/article/10.1007/s10055-022-00638-2) --- # CEFR Levels Explained: What A1 to C2 Mean for Speaking (and How Long Each Takes) URL: https://oravitavr.com/blog/cefr-levels-explained/ > **Quick answer:** The CEFR (Common European Framework of Reference for Languages) describes language ability on six levels: A1 and A2 (basic user), B1 and B2 (independent user), and C1 and C2 (proficient user). Cambridge English estimates about 350–400 guided learning hours to reach B1 and 500–600 hours to reach B2, though the time varies by language and learner. If you have ever seen "B2 required" in a job advert or been placed in an "A2 class", you have met the CEFR. It is the most widely used scale for describing what someone can do in a language. This guide explains each level with a focus on speaking, because that is where the levels feel most real. ## What is the CEFR? The Common European Framework of Reference for Languages (CEFR) is a framework published by the Council of Europe that describes language ability in terms of what learners *can do*, not just what they know. It applies to any language, so a B1 in Spanish means the same as a B1 in Japanese. In 2020 the Council of Europe released the *CEFR Companion Volume*, which expanded the descriptors with areas such as mediation (helping others understand) and online interaction ([Council of Europe, 2020](https://rm.coe.int/cefr-companion-volume-with-new-descriptors-2020/16809ea0d4)). ## The 6 CEFR levels at a glance | Level | Name | In one sentence (speaking) | |---|---|---| | A1 | Breakthrough | You can introduce yourself and ask simple questions if the other person speaks slowly. | | A2 | Waystage | You can handle routine tasks such as shopping, ordering food and asking for directions. | | B1 | Threshold | You can deal with most situations while travelling and talk about experiences, plans and opinions. | | B2 | Vantage | You can hold a natural conversation with native speakers and argue a point of view. | | C1 | Effective operational proficiency | You can speak fluently and flexibly at work and in study, without searching for words. | | C2 | Mastery | You can express yourself precisely, spontaneously and with fine shades of meaning. | ## What can you say at each CEFR level? ### A1: Breakthrough At A1 you use familiar everyday expressions and very basic phrases. You can introduce yourself, say where you live and ask someone how they are, as long as they speak slowly and help you. **Example:** "Hello, I'm Anna. I live in Zurich. And you?" **Practise by:** Repeating short dialogues out loud and shadowing simple audio. ### A2: Waystage At A2 you can communicate in simple, routine tasks. You can talk about your background, your family and your immediate needs, and handle short exchanges in shops, cafés and stations. **Example:** "Could I have a coffee and a croissant, please? How much is that?" **Practise by:** Rehearsing everyday situations: ordering, buying tickets, asking for directions. ### B1: Threshold B1 is the point where you become an *independent user*. You can deal with most situations while travelling, describe experiences and events, and briefly explain your opinions and plans. **Example:** "Last summer I went to Lisbon. I loved the old town, but it was very hot, so next time I'd go in spring." **Practise by:** Unscripted conversation. B1 learners know enough to talk, but often freeze when the conversation goes somewhere unexpected. ### B2: Vantage At B2 you can interact with native speakers with a degree of fluency and spontaneity that makes conversation comfortable for both sides. You can explain a viewpoint and give the pros and cons of different options. **Example:** "I see your point, but I think remote work only works if teams meet in person regularly." **Practise by:** Discussions and debates on topics you care about, and role-plays of work situations. ### C1: Effective operational proficiency At C1 you speak fluently without obviously searching for expressions, and you use language flexibly for social, academic and professional purposes. **Practise by:** Presentations, negotiations and conversations where register and tone matter. ### C2: Mastery At C2 you understand virtually everything you hear and can express yourself spontaneously, very fluently and precisely, even in complex situations. **Practise by:** Varied, demanding conversations with feedback on nuance, idiom and style. ## How long does it take to reach each CEFR level? Cambridge English gives these approximate cumulative guided learning hours for English, starting from zero ([Cambridge English](https://support.cambridgeenglish.org/hc/en-gb/articles/202838506-Guided-learning-hours)): | Level | Guided learning hours (cumulative) | |---|---| | A2 | 180–200 | | B1 | 350–400 | | B2 | 500–600 | | C1 | 700–800 | | C2 | 1,000–1,200 | Cambridge stresses these are guidelines only. Your background, study intensity and practice outside lessons all change the numbers. The language also matters. For English speakers, the US Foreign Service Institute estimates that reaching professional working proficiency takes about 24–30 weeks (600–750 class hours) for languages such as Spanish, French and Italian, about 36 weeks (900 hours) for German, and about 88 weeks (2,200 hours) for Japanese, Korean, Chinese and Arabic ([U.S. Department of State](https://www.state.gov/foreign-language-training/)). ## Why do learners get stuck at B1? Many learners describe a "B1 plateau": they understand a lot, but speaking does not get easier. The usual reason is that B1 is where *spontaneous* speaking becomes the main skill, and spontaneous speaking only improves with a lot of real conversation. Grammar exercises and vocabulary apps stop being enough. The fix is volume: many short, unscripted conversations on varied topics, ideally with feedback. If a lack of conversation partners is the problem, see [how to practise speaking a language alone](/blog/practise-speaking-a-language-alone/). ## How do you find out your CEFR level? - **Official exams** such as Cambridge English, DELF/DALF for French, Goethe-Zertifikat for German and DELE for Spanish give a certified level. - **Placement tests** at language schools give a quick estimate. - **Self-assessment** with the CEFR "can do" statements is a free first check: be honest about what you can do *in conversation*, not only on paper. ## Key takeaways - The CEFR has six levels, from A1 (beginner) to C2 (mastery), grouped as basic, independent and proficient users. - B1 is the turning point where spontaneous speaking becomes the main skill. - Reaching B2 in English takes roughly 500–600 guided learning hours, according to Cambridge English. - Speaking practice should match your level: repetition at A1–A2, free conversation from B1. ## How Oravita uses the CEFR [Oravita](/) maps every scene and conversation to the CEFR, from A1 to C2. The AI teacher adjusts its pace, vocabulary and topics to your level as you speak, so an A2 learner orders a coffee while a B2 learner debates the future of city centres in the same street. Oravita is in beta now. [Join the beta](/#beta). ## Frequently asked questions ### What are the 6 CEFR levels? The six CEFR levels are A1 (Breakthrough), A2 (Waystage), B1 (Threshold), B2 (Vantage), C1 (Effective operational proficiency) and C2 (Mastery). ### Is B2 considered fluent? B2 is often described as conversationally fluent: you can interact with native speakers naturally and without strain on either side. Many employers and universities ask for B2 or C1. ### How many hours does it take to get from B1 to B2? Based on Cambridge English guidelines (about 350–400 hours for B1 and 500–600 hours for B2), moving from B1 to B2 in English takes roughly another 150–200 guided learning hours, though this varies widely between learners. ### What CEFR level do I need for a job? It depends on the role. Many jobs that use a language daily ask for B2, while roles involving negotiation, writing or client contact often require C1. ### Can I skip a CEFR level? You cannot skip the skills a level describes, but you can move through levels faster with intensive practice, especially speaking practice. ## Sources - Council of Europe (2020). *Common European Framework of Reference for Languages: Companion Volume*. [Link](https://rm.coe.int/cefr-companion-volume-with-new-descriptors-2020/16809ea0d4) - Cambridge English. Guided learning hours. [Link](https://support.cambridgeenglish.org/hc/en-gb/articles/202838506-Guided-learning-hours) - U.S. Department of State, Foreign Service Institute. Foreign language training. [Link](https://www.state.gov/foreign-language-training/) --- # Speaking Practice at Scale: A Guide for Language Schools and Companies URL: https://oravitavr.com/blog/language-training-for-schools-and-companies/ > **Quick answer:** Language schools and companies can give every learner more speaking practice by adding AI-powered conversation, ideally in immersive VR, between lessons. It does not replace teachers; it multiplies speaking time, lets learners rehearse real situations at their CEFR level, and gives organisations data on progress that classroom time alone cannot provide. Whether you run a language school, a university programme or learning and development at a company, you face the same problem: learners need to speak, and there is never enough time for everyone to do it. This guide looks at why that gap matters and how to close it. ## Why is speaking the weak spot of most language programmes? Because speaking time is shared. In a group lesson, one teacher divides attention across many learners. A large German study of English lessons, which recorded classes in 219 schools with around 11,000 students, found that students as a whole group spoke for less than a quarter of lesson time ([DESI-Konsortium, 2008](https://extemporeapp.com/blog/how-much-do-students-speak-in-the-target-language)). Speaking is also the skill learners feel most anxious about. In the study that defined foreign language anxiety, 49% of students said they panicked when asked to speak without preparation ([Horwitz, Horwitz & Cope, 1986](https://www.ling.ohio-state.edu/research/groups/lacqueys/readings/horowitz.pdf)). Little practice plus high anxiety is why many learners leave courses able to read and write well but unable to hold a conversation. ## What does poor speaking ability cost organisations? For companies, the cost shows up in business results. In an Economist Intelligence Unit survey of 572 executives, 49% said communication misunderstandings had stood in the way of a major international business deal, and 47% said their company did not provide enough training in language and communication skills ([Economist Intelligence Unit, 2012](https://www.ibs-b.hu/documents/278/Economist_Intelligence_Unit_-_Competing_across_borders_2012.pdf)). Nine in ten believed better cross-border communication would improve profit, revenue and market share. For schools and universities, the cost is learner outcomes and satisfaction. Learners who cannot speak after a course rarely recommend it. ## How can AI and VR add speaking practice at scale? AI conversation partners and immersive VR solve the time problem from different angles: - **AI conversation** gives every learner a patient partner who is available 24/7 and responds in real time. A 2025 meta-analysis of 28 studies found a positive overall effect of chatbots on language learning performance (g = 0.48) ([Wang et al., 2025](https://journals.sagepub.com/doi/abs/10.3102/00346543241255621)). - **VR** places that conversation in a realistic setting, such as a client meeting, a hotel reception or a hospital ward, so learners rehearse the exact situations they will face. Practising speaking in high-immersion VR has been linked to significantly lower foreign language anxiety than the same practice on Zoom ([Kaplan-Rakowski & Gruber, 2023](https://link.springer.com/article/10.1186/s40561-023-00263-9)). - **Together,** they let each learner speak for as long as they need, in context, without waiting for their turn. Immersive training also scales well economically. In PwC's study of VR soft-skills training, VR reached cost parity with classroom training at 375 learners and became 52% cheaper than classroom training at 3,000 learners ([PwC, 2020](https://www.pwc.co.uk/services/technology/immersive-technologies/study-into-vr-training-effectiveness.html)). ## Does AI replace language teachers? No. The most effective setup is blended: teachers introduce structures, give feedback and keep learners motivated, while AI handles the high-volume speaking practice that a single teacher cannot provide to a whole group. Teachers get learners who arrive at class having already spoken, and data on where each learner struggles. ## What should you look for in a speaking practice solution? | Criterion | Why it matters | |---|---| | Unscripted, real-time conversation | Learners must practise responding to the unexpected, not reciting lines | | CEFR alignment from A1 to C2 | Fits your existing levels, classes and certificates | | Scenario relevance | Learners should rehearse situations from their real work or study | | Language and accent coverage | Teams and classes often need more than one language or regional variant | | Progress reporting | Shows learners, teachers and managers what is improving | | Onboarding and certificates | Makes adoption easy and gives learners recognition | | Hardware flexibility | Supports devices you already own or provides them | | Scientific validation | Separates proven approaches from marketing claims | ## How do you measure the impact of speaking practice? Agree on a few simple measures before you start: 1. **Speaking time per learner per week,** compared with before. 2. **CEFR progression** on speaking, assessed at the start and end of a period. 3. **Learner confidence,** using a short self-assessment survey. 4. **Business or academic outcomes,** such as fewer escalations in client calls, better exam results or higher course completion. A four-to-eight-week pilot with one class or team is usually enough to see whether speaking time and confidence move. ## Key takeaways - Group lessons leave most learners with very little speaking time. - Weak speaking skills cost companies deals and cost schools learner satisfaction. - AI conversation, especially in VR, multiplies speaking time without replacing teachers. - Choose a solution with unscripted conversation, CEFR alignment and progress reporting, and start with a short pilot. ## Oravita for organisations [Oravita](/) is built for language schools, universities and companies. Learners speak with an AI teacher in real time, inside generated real-life environments, at their CEFR level from A1 to C2, in any language and accent. Organisations get onboarding, certificates and progress reporting, with hardware supplied. Oravita's prototype is IEEE-published and its research is backed by Innosuisse, the Swiss Innovation Agency. Oravita is in beta now. [Talk to us about a pilot](/contact/). Related: [CEFR levels explained](/blog/cefr-levels-explained/) · [Fear of speaking a foreign language](/blog/fear-of-speaking-a-foreign-language/) ## Frequently asked questions ### How can companies improve employees' speaking skills in a foreign language? Give employees frequent, low-pressure speaking practice in situations from their real work, for example with an AI conversation partner or VR role-plays, and combine it with teacher-led sessions and clear CEFR goals. ### What is the ROI of corporate language training? It depends on the role, but poor communication has a measurable cost: in an Economist Intelligence Unit survey, 49% of executives said communication misunderstandings had stood in the way of a major international deal. ### Can AI replace language teachers? No. AI is best at providing high-volume speaking practice and feedback between lessons. Teachers remain essential for structure, explanation and motivation. ### How do language schools use VR? Language schools use VR to give learners realistic speaking practice between classes, for example ordering at a café, checking into a hotel or holding a job interview, so each learner speaks far more than group lessons allow. ### How long should a language training pilot run? Four to eight weeks with one class or team is usually enough to measure changes in speaking time, confidence and early CEFR progress. ## Sources - DESI-Konsortium (2008). *Unterricht und Kompetenzerwerb in Deutsch und Englisch.* Summarised in [Extempore](https://extemporeapp.com/blog/how-much-do-students-speak-in-the-target-language). - Horwitz, E. K., Horwitz, M. B. & Cope, J. (1986). Foreign language classroom anxiety. *The Modern Language Journal*. [Link](https://www.ling.ohio-state.edu/research/groups/lacqueys/readings/horowitz.pdf) - Economist Intelligence Unit (2012). *Competing across borders: How cultural and communication barriers affect business.* [Link](https://www.ibs-b.hu/documents/278/Economist_Intelligence_Unit_-_Competing_across_borders_2012.pdf) - Wang, F. et al. (2025). A meta-analysis of chatbot-assisted language learning. *Review of Educational Research*. [Link](https://journals.sagepub.com/doi/abs/10.3102/00346543241255621) - Kaplan-Rakowski, R. & Gruber, A. (2023). The impact of high-immersion virtual reality on foreign language anxiety. *Smart Learning Environments*. [Link](https://link.springer.com/article/10.1186/s40561-023-00263-9) - PwC (2020). The effectiveness of virtual reality soft skills training in the enterprise. [Link](https://www.pwc.co.uk/services/technology/immersive-technologies/study-into-vr-training-effectiveness.html)