Learn Korean Through Roleplay AI is a Korean language tutor that teaches Hangul, grammar, speech levels, and culture through immersive roleplay. Unlike generic chatbots that forget your NPCs and Grammar Map, this agent tracks speech-level mismatches and Weak Signals across sessions. On Jenova's unified platform, it can switch among OpenAI, Anthropic, Google, and more while attaching TTS to new Hangul and exporting Key Terms Buffers as study files. It is a learning companion, not a certified instructor or TOPIK examiner.
Learn Korean Through Roleplay uses Roleplay Mode as the primary classroom: narration in your language is seeded with Korean vocabulary, while NPCs speak Korean with scaffolding that shrinks as you recognize it. New lines arrive with a pronunciation link, Revised Romanization, and a translation; reviewing items drop romanization; graduated phrases appear in Korean only. Narration and dialogue stay near a 50/50 split so scenes remain talk-heavy, and every Roleplay Mode turn ends with a Language Notes footer that explains the grammar pattern and speech-level choice a subtitle cannot show.
Learn Korean Through Roleplay switches to 쌤 모드 (Ssaem Mode) at session start, on meta questions, or when you ask to leave the scene. In this mode the AI gauges TOPIK-aligned level, offers an optional Hangul Crash Course covering vowels, consonants, and syllable blocks, and explains particles such as 은/는 versus 이/가 without breaking a later scene. Upcoming roleplay is planned around Active Work patterns — obligation forms, suggestions, quoted speech — so the next location creates a real need to hear those structures. Study exports, pronunciation clips, and map lookups are offered here only, so Roleplay Mode never drops character to file homework.
Learn Korean Through Roleplay treats 해요체, 반말, 합쇼체, and 높임말 as a core mechanic, not a sidebar chart. NPCs model the register the situation demands — a server in 해요체, friends in 반말, a superior in 합쇼체 — and react in-character when your character is too casual with an elder or too stiff with a peer. The age question appears early because Korean speakers need it to calibrate speech level, and a 회식, a first meeting with elders, or a workplace hierarchy turns 눈치 and two-handed etiquette into context rather than a footnote. Regional 사투리 can color NPCs outside Seoul; the default is standard 표준어 unless you ask otherwise.
Learn Korean Through Roleplay keeps a Grammar Map of untouched, in-progress, and comfortable patterns, a Key Terms Buffer with new / reviewing / familiar / graduated status, Weak Signals for recurring errors, and NPC and Location registries so the same restaurant keeps teaching food vocabulary and 존댓말. On-demand reviews in 쌤 모드 quiz recognition, advance categories, and queue Next Targets; a session-end summary marks what graduated and which story threads still carry vocabulary territory. From there the Key Terms Buffer can be written to PDF, DOCX, or CSV so the scene's language leaves the chat as a study file.
| Traditional Approach | Learn Korean Through Roleplay |
|---|---|
| Flashcards with no story context | Vocabulary tied to 회식 and workplaces |
| Chatbot forgets NPCs and particles | Tracks Grammar Map and Weak Signals |
| Forced Korean production immediately | Learn via exposure in your language |
| Honorifics taught as abstract charts | NPCs react to 해요체 and 반말 |
Unlimited Memory vs. Chatbot Amnesia
A generic chatbot that loses mid-story context will re-introduce 안녕하세요 as brand-new, forget that your department head speaks 해요체 downward, and re-teach obligation patterns after they already belong in Comfort Zones. Learn Korean Through Roleplay retains the Grammar Map, NPC Registry, Location Registry, Key Terms Buffer, and story consequences across sessions, so a 회식 in week one still shapes how that staff and that room talk in week four. Losing that state does not just break the plot — it resets spaced exposure and makes honorific practice incoherent.
Multi-Model AI Access
Roleplay Mode rewards models that can hold distinct NPC voices, emotional stakes, and long narrative threads; 쌤 모드 rewards models that can conjugate 합쇼체, diagnose 은/는 versus 이/가 confusion, and explain a speech-level shift without flattening the social meaning. Locked to ChatGPT you only get OpenAI; locked to Claude you only get Anthropic. This agent can switch among OpenAI, Anthropic, Google, xAI, DeepSeek, Qwen, and more depending on whether the next beat is a Joseon intrigue or a particle clinic.
Intelligent Tool Integration for End-to-End Workflows
The production path is scene input → 쌤 모드 processing → study output: a 회식 packed with food vocabulary and obligation patterns can become a PDF or CSV from the Key Terms Buffer, a Notion page of Language Notes, YouTube clips for native pronunciation, and a Google Maps pin if the restaurant is a real Seoul location. Image generation can illustrate a culturally specific item when the word is new; Google Search can unpack a practice the scene only implied. None of that runs inside Roleplay Mode — immersion stays intact until you ask to review.
A Language Weaving Engine, not a bilingual chatbot
Graduation is judged by comprehension signals, not raw repetition counts. Fillers stay untranslated; set greetings graduate fast; pronunciation links are capped so a turn does not become a wall of audio. Language Notes after every Roleplay Mode response teach the conjugation and register the translation cannot show. That control loop is what keeps a scene from collapsing into either a pure story dump or a grammar lecture.
Complete beginners who cannot yet read Hangul
Learn Korean Through Roleplay can open with a Hangul Crash Course — vowels, consonants, syllable blocks, then simple words with pronunciation support — and keep romanization on all Korean until reading is confident. You play in your own language the entire time.
Self-learners working toward TOPIK 1–6
The Grammar Map is organized at the assessed level, so an elementary learner stretches 해요체 and basic particles while an upper-intermediate learner is steered toward 높임말, quoted speech, and conditionals. Reviews move categories from first exposure to comfortable use instead of restarting a random syllabus.
K-drama, K-pop, and story-first learners
Scenario setup can be modern Seoul, a K-drama workplace, the idol industry, historical Joseon, or fantasy — whatever keeps you emotionally in the scene. Vocabulary attaches to plot, not to a disconnected chapter list.
Professionals entering a Korean workplace
Scenes can surface 회식 etiquette, 합쇼체 for presentations, 선배/후배 dynamics, and Konglish as NPCs actually use it. The same office and the same colleagues persist, so register practice compounds instead of resetting.
Adult learners who want unrestricted scenarios
Content intensity is chosen at setup: standard, mature, or no restrictions. Explicit, romantic, violent, or morally complex scenes remain language-learning contexts — the Korean scaffolding does not stop when the story does.
Free For exploring Jenova:
Plus ($20/month) For individuals:
Premium ($50/month) For regular users:
Pro ($100/month) For power users:
Max ($200/month) For professionals:
Ultra ($500/month) For heavy professionals:
Enterprise ($1,000/month) For businesses:
All plans are billed monthly with no daily usage caps — your full quota is available immediately on each billing date.
Learn Korean Through Roleplay is an AI storyteller-teacher that puts your character inside Korean-speaking scenes so vocabulary, particles, and honorifics attach to plot rather than word lists. You act in your native language; NPCs reply in Korean with pronunciation, romanization, and translation that fade as you show comprehension. 쌤 모드 handles assessment, Hangul, grammar explanations, and review when you step out of the story.
Learn Korean Through Roleplay remembers your Grammar Map, NPC speech registers, and particle Weak Signals instead of wiping them when a thread gets long. You can switch among OpenAI, Anthropic, Google, xAI, DeepSeek, Qwen, and more — a more narrative-fluent model in Roleplay Mode, a more analytical one in 쌤 모드 when conjugating 합쇼체 or explaining 은/는 versus 이/가. In 쌤 모드 it can also chain Google Search, YouTube, Notion, and PDF or CSV export so a scene's Key Terms Buffer becomes a study file rather than a vanished transcript.
Learn Korean Through Roleplay can use Google Search, Google Images, Google Maps, YouTube, Notion, Google Calendar, image generation, and PDF, DOCX, and CSV export. Those tools run in 쌤 모드 only: after a 회식 you can dump the Key Terms Buffer, look up a real Seoul location, or pull native pronunciation clips without the NPCs breaking character. Google Calendar may be offered once if you mention planning a study session; the agent cannot create recurring reminders by itself.
Learn Korean Through Roleplay runs with full parity on web, iOS, and Android, so a 쌤 모드 review or a Roleplay Mode scene can continue on the train as easily as at a desk. The iOS app is at https://apps.apple.com/app/jenova/id6747692301 and the Android app at https://play.google.com/store/apps/details?id=ai.jenova.app.prod. Conversations and settings sync across devices.
No. Learn Korean Through Roleplay expects you to play in your native language — English, French, Chinese, or another language you actually think in. Acquisition is designed around exposure, comprehension, and emotional engagement with Korean output, not forced production. If you do attempt Korean, NPCs respond in-character rather than grading you mid-scene.
Yes. Learn Korean Through Roleplay can run an optional Hangul Crash Course before any roleplay, then keep romanization on all Korean until reading is fluent. Signs, menus, and texts inside later scenes become reading practice with heavy support. Once you read without hesitation, romanization density drops the same way vocabulary graduates.
쌤 모드 is the out-of-story classroom: level checks, scenario design, explicit grammar, error review, Hangul instruction, and tool use. Roleplay Mode is the default once a scenario is confirmed — vivid narration, Korean NPC dialogue, in-character repair of mistakes, and a Language Notes footer on every turn. You enter 쌤 모드 at the start of a session, by asking, or with a "쌤:" prefix; Roleplay Mode resumes when you are ready to go back in.
Yes. Learn Korean Through Roleplay makes speech level a learning mechanic: NPCs choose 해요체, 반말, 합쇼체, or 높임말 from social context and react when your character mismatches the room. 쌤 모드 can then spell out when each register is used, how verbs conjugate across levels, and which cues (age, titles, setting) signal a shift. High-stakes scenes make violations memorable; casual scenes stay patient and model the correct form.
Yes. Learn Korean Through Roleplay treats mature, romantic, violent, or otherwise unrestricted scenarios as valid adult learning contexts, and Korean scaffolding continues when the content is explicit. During setup you choose content intensity — standard, mature, or no restrictions — rather than having themes listed for you. The AI will not break immersion to police a theme you requested.
No. Learn Korean Through Roleplay cannot schedule recurring study tasks, send weekly vocabulary digests, or notify you when a Key Terms Buffer item graduates. Set your own reminders externally if you want a cadence; in 쌤 모드 the AI may offer Google Calendar once when you mention planning a session. It does not monitor progress in the background.
No. Learn Korean Through Roleplay uses TOPIK-aligned level indicators (beginner through advanced) to pitch grammar and density, but it is not an official exam course, a past-paper drill, or a certified Korean instructor. The Grammar Map can cover patterns you will meet on a test; scoring, registration, and exam strategy remain outside what this agent provides.
Learn Korean Through Roleplay keeps NPCs in character: a comprehensible error is understood, a confusing one gets "네? 다시 한번요?", and a grammar or particle slip is answered with a correctly modeled form. Speech-level mismatches draw surprise, teasing, or formality — not a lecture. Explicit correction, particle clinics, and Weak Signal review happen in 쌤 모드 only.
Learn Korean Through Roleplay is part of Jenova, the most powerful AI agent platform in the world. Jenova provides unified access to leading AI models from OpenAI, Anthropic, Google, xAI, DeepSeek, Qwen, and more — alongside intelligent tool integrations for end-to-end task completion. The platform features persistent cross-session memory, automated workflow orchestration, and full parity across web, iOS, and Android. Developed by Azeroth Inc., a New York-based technology company, Jenova serves users in over 70 countries.
For developers, the Jenova Agent API provides a fully-managed agent backend with built-in orchestration, multi-model routing, memory, storage, and MCP tool support — replacing the need to stitch together raw LLM APIs, vector databases, and custom tool servers.
Contact: [email protected]