Our innovative language app utilizes state-of-the-art AI and mixed reality to ​redefine language learning, converting your everyday surroundings into an ​interactive learning environment. Imagine turning your living room into a ​vibrant educational space where ordinary items become language tools. With ​our advanced AI vision models and object detection, the app links new ​vocabulary to familiar objects, crafting a personal and intuitive spatial ​memory system.

A standout feature of the app is its ability to understand context and adapt in ​real-time to your environment. It stays updated as you rearrange or introduce ​new items, ensuring your learning experience remains dynamic and engaging. ​Picture learning the word "book" when picking up a novel today, and "set the ​table" while arranging dinner tomorrow.

What distinguishes our app is its power to make the ordinary extraordinary. By ​associating new language concepts with known objects and settings, we make ​learning not just intuitive but memorable. This approach enhances vocabulary ​retention and recall, ensuring long-term mastery.

Ultimately, our app is more than just a vocabulary builder—it's an AI-driven ​transformation that turns your home into a personalized, interactive ​language school. By integrating language learning into your daily life and ​simulating real-world interactions, our app makes mastering a new language ​both effective and exciting.

Jargon understands your environment, teaching you relevant vocabulary as you interact with ​your surroundings. Learn naturally by connecting new words to real objects, making the ​learning process feel effortless and intuitive.

Your learning evolves with you. As you change your surroundings, Jargon adapts instantly—​keeping lessons relevant to what you’re doing. Whether cooking, cleaning, or relaxing, Jargon ​matches your activities to enhance retention.

Learning becomes part of your daily routine. With Jargon, you’re not just memorizing words—​you’re living them. By connecting language to everyday actions, Jargon ensures long-lasting ​mastery in a way that’s both fun and effective.

Designed for beginners, Jargon offers a fun, stress-free experience. No pressure, just playful ​lessons that help you learn naturally and enjoyably.

If you’re a fan of mobile language apps and curious about immersive learning, Jargon brings the ​ease of app-based learning to mixed reality—offering an experience that makes learning intuitive ​and exciting.

Jargon is for those excited to explore mixed reality while learning a new language. Developed for ​Meta Quest, it turns your everyday environment into an interactive language playground.

The Jargon experience is designed to be immersive, playful, and highly interactive. By blending ​real-world objects with mixed reality, Jargon creates an engaging environment where players a​cquire new language skills through interaction. The application is built around a stress-fr​ee and whimsical experience, encouraging exploration and curiosity without traditional le​arning pres​sures.

Jargon, the AI assistant, is a cute, owl-like robot that serves as the player’s friendly guide ​throughout their language learning journey. Interacting with Jargon is primarily done through ​voice commands, making it feel intuitive and hands-free. Jargon’s role is to facilitate learning by ​explaining new concepts, providing challenges, and offering encouragement—making learning ​feel like an exciting adventure rather than a formal lesson.

The wristband interface acts as a secondary, supportive UI for displaying essential information ​such as player level, game modes, and language selections. While the main interactions happen ​via voice commands with Jargon, the wristband provides a traditional interface for those times ​when players need a quick visual overview or want to make manual selections. It offers a ​complementary way to interact without interrupting the immersive experience.

Jargon’s design focuses on being approachable and fun. The owl-like robot embodies a mix of ​technology and nature, rendered in light pastel colors that evoke a sense of calm and ​friendliness. The surrounding environment is vibrant and full of interactive elements, ​encouraging players to explore and interact with real-world and virtual objects. Together, Jargon ​and the environment form a cohesive, playful atmosphere that turns every learning moment into ​an engaging adventure.

Sound effects and music are light, playful, and designed to support the whimsical nature of the ​app. They provide positive auditory feedback, reinforcing learning moments and keeping players ​engaged in a cheerful way.

• Voice Interaction: The primary method of interacting with Jargon is through voice ​commands. Players ers learn vocabulary, receive instructions, and complete challenges using ​their voice, which makes the experience feel natural and keeps both hands free.
  

• Hand Tracking and Wristband UI: Players can interact with their surroundings through ​hand gestures, and use the wristband interface to access more traditional UI elements. The ​combination of hand tracking and wristband input ensures players have full control, from an ​immersive, voice-driven experience to a more conventional interface for specific ​interactions.
  

Jargon guides players through their language learning journey. The experience begins simply—​labeling objects in the environment to establish foundational vocabulary. As players progress, ​they build on this foundation, forming phrases, sentences, and eventually more complex ​dialogues. Jargon supports players along the way, providing challenges to deepen their ​understanding.

The journey is dynamic, shaped by the player’s environment. By adding or removing items from ​the playspace, or even moving to different rooms or locations, players experience a different ​learning path. Each new environment introduces new vocabulary and phrases, allowing the ​player to shape their own learning experience. The more varied the surroundings, the richer the ​language journey becomes, with Jargon always adapting to guide and support growth.

The core goal of Jargon is for players to learn a new language through voice-driven exploration ​and interaction. Players use their acquired skills to complete challenges involving pronunciation, ​sentence construction, and contextual vocabulary use—all of which makes learning an active and ​rewarding process.

In the initial three months of development, the team will focus on creating a functional mechanical proof of concept, ​establishing the core game loop. Upon opening the app, players will be greeted by a home screen. From there, they will be ​prompted to grant necessary permissions, such as microphone input and passthrough access, and select the target ​language they wish to learn. Once permissions are granted, the app, Jargon, will collect playspace information to begin ​automatically identifying and labeling objects within the player's environment.

In Level 1, players will engage in word translation exercises, learning the names of objects in their playspace in the ​chosen language.

Progressing to Level 2, players will be challenged to construct and articulate simple sentences. Successful ​pronunciation will allow Jargon to perform actions within the playspace, enhancing the interactive and immersive ​learning experience.

After several months of development, the team is ready to implement our production feedback loop. We will ​gather player feedback and iterate on our designs to enhance the experience for our players. By listening to our c​ommunity and prioritizing the most requested features, we aim to provide more value and enjoyment. A​dditionally, we have our own exciting feature ideas that we're eager to introduce.

Role-Playing: This feature transports players to another world, ​allowing them to take on different personas. Players will need to ​listen attentively and act out the scenarios presented to them.

Scavenger Hunts: Players will receive a word in another language and ​must find the corresponding virtual object hidden in their play space. ​Be prepared for a few playful tricks along the way!

Escape Rooms: Team up with a character named Jargon and use another ​language to communicate and solve puzzles to escape a room within the ​allotted time.

Our goal is to create an experience that players love, and your ​feedback is crucial in achieving this. We look forward to hearing your ​thoughts and making this journey together.

Jargon first collects data from both the player's voice interactions and the mixed reality environment. This information ​is compiled into a prompt, which includes the message history, a system prompt, previously learned words, and any ​available scene functions.

Jargon processes these inputs to determine the best response.

Once a decision is made, Jargon can execute several actions:

• Function Calls: If Jargon decides to call one or more functions, the app executes these to perform specific tasks.
  
• Structured Output: If Jargon generates a structured response, such as a lesson plan, the app updates the player’s ​interface accordingly.
  
• Text Response: If Jargon responds with text, it is converted into speech using Jargon's voice and played within the app
  

Jargon is capable of employing all these response types simultaneously in a single interaction, ensuring a dynamic and ​engaging player experience.

In the first few months of development, Jargon will be free to use as we work towards proving ​the concept and building a vertical slice of the app. During this phase, our focus is on gathering ​valuable player feedback to refine and improve Jargon for a full production release.

As players interact with Jargon, the app will generatively produce new language content, whic​h incurs a small cost for each new generation. To mitigate this, we plan to cache generated content​, allowing it to be reused across different features of the app, such as knowledge tests, reusabl​e lesson plans, and a growing word bank. This caching approach reduces the need for repeate​d content generation, saving costs over time while still providing players with a rich experien​ce.

With an increasing player base and the implementation of database solutions to store generated ​content, these generation costs will gradually increase. To manage this, we plan to implement ​one of the following monetization models:

• Credit System: Players purchase credits that are used whenever new content is generated by ​Jargon. This provides a flexible, pay-as-you-go approach.
  
• Subscription System: A monthly subscription that offers unlimited content generation for ​the duration of the subscription. This appeals to players who prefer consistent access without ​managing credits.
  
• Hybrid System: A combination of credits and subscription, where players pay a monthly fee ​and receive a set amount of credits each month, providing a balance of flexibility and ​predictability.
  

Players will use credits primarily when new content is generated, such as exploring a new ​environment and learning new vocabulary or phrases, generating new lesson plans, or ​advancing to conversational practice with Jargon. Since previously generated content will be ​cached and reused, credit usage will be limited to situations where unique, new content is ​required, optimizing the value preceive.

This approach ensures that players pay for new and personalized content while enjoying ​ongoing value from previously generated learning materials, creating a sustainable model for ​both the player and the development of Jargon.

Over the past few years, AI models have shown a clear trend of ​becoming both more powerful and less expensive. From GPT-3 to ​GPT-4o and Llama 3.1, these advancements have brought more ​capabilities at reduced costs per input/output token. As we continue ​to use these evolving services, we expect that the costs for inference ​and fine-tuning will decrease over time. This trend will enable us to ​implement cutting-edge models to enhance Jargon while also ​making it more viable to fine-tune or quantize models that are ​already “good enough” for our specific use cases.

In the coming years, we also anticipate support for even more ​languages, further expanding Jargon’s capabilities. This will open ​up opportunities to introduce advanced, specialized content, ​including niche areas like technical jargon, business jargon, medical ​jargon, etc…—catering even to native speakers who want to deepen ​their expertise.

If the costs of using these models do not decrease as expected, the ​advancements in computing devices may provide a viable solution. ​Smaller, “good enough” models could potentially run directly on the ​client’s device using Unity Sentis (Unity's Neural Inferencing Engine), ​which is currently in beta. This approach could significantly reduce ​or even eliminate inference costs, while also improving latency by ​avoiding reliance on server communication or external APIs. ​Running models locally would allow us to freeze them when no ​updates are needed, reducing maintenance costs and shielding us ​from disruptions caused by changes or the discontinuation of third-​party services.

Ultimately, Jargon stands to benefit from the ongoing improvements ​in AI model capabilities. As models become more powerful, respond ​more human-like, and become more cost-effective, Jargon will be ​able to focus on what it does best—delivering an engaging, contextual ​learning experience. These advances enable us to improve ​continuously, while letting Jargon grow alongside the best AI ​technologies available.