Chinese Chess Coach AI is an adaptive Xiangqi tutor that teaches piece mechanics, openings, tactics, and endgames through personalized coaching at every skill level. Unlike generic chatbots that treat Xiangqi as chess with different pieces, this agent coaches screen management, palace integrity, and named mates as Xiangqi theory. Built on Jenova's unified platform, it combines models from OpenAI, Anthropic, Google, and more so a rules lesson and a Central Cannon post-mortem are not stuck on a single vendor.
Chinese Chess Coach infers your level from how you talk — how the horse moves, Screen Horse by name, ECCO codes, or a pasted FEN — then confirms the read and adjusts. The work is whole-board thinking: piece coordination, tempo, screens, and structural integrity, not isolated moves.
Chinese Chess Coach teaches Xiangqi rules in a teachable order: board, river, and palace first, then pieces from chariot and soldier through general, advisor, horse, elephant, and cannon. Check, checkmate, and stalemate-as-loss wait until the pieces make sense; flying general and forbidden repetitions wait until you can already play a legal game. Chess converts hear the false friends named early — the horse blocked at the knee (蹩马腿), the cannon that captures only by jumping a screen, the general confined to the palace, pieces sitting on intersections (9 files × 10 ranks).
Chinese Chess Coach teaches openings as systems with plans you can still use when the opponent leaves book on move eight. Central Cannon versus Screen Horse Defense is coached as the main clash; Same Direction Cannons are treated as sharp; Angel Pawn and Elephant Opening as quieter systems that transpose heavily. You learn why chariots develop first, why the fifth file is contested terrain, when a ranked chariot is more useful than a filed chariot, and when to trade a cannon that is about to devalue.
Chinese Chess Coach drills named mates and tactics as patterns you answer first, then have explained — so the idea is yours instead of a spoiler. One-move mates build into multi-move combinations, defensive saves, and positions where your own piece has become a screen against you. The vocabulary is the vocabulary you will actually use over the board.
Chinese Chess Coach classifies endings the way Xiangqi literature already maps them: theoretical win, theoretical draw, positional win, positional draw. The first question is always the defender’s guard — two advisors plus two elephants, full guard, changes what extra material can force. A chariot versus full guard is usually a draw; chariot-plus-horse is a win with best play, which is why this agent wants the target ending chosen before middlegame trades start.
Chinese Chess Coach walks a pasted FEN, WXF scoresheet, Chinese notation, or spoken recount the way a coach does, not the way a graph does: confirm opening, middlegame, or time trouble; assess material and structure; list candidate moves; recommend one; name the principle. Full-game reviews flag turning points and cross-move habits — leaving a horse blockable, trading cannons while screens still exist, steering into a chariot-versus-full-guard draw when a remaining horse still made a theoretical win. Practical moves at your strength are separated from engine-only shots, then a follow-up drill targets the pattern you missed.
| Traditional Approach | Chinese Chess Coach |
|---|---|
| Static lessons ignore your hanging cannons | Drills built from your last post-mortem |
| Chess stalemate habits cost winning games | Stalemate-as-loss taught as a weapon |
| Engine evals without a named principle | Every recommendation names the idea |
| Memorized lines collapse off book | Opening systems with playable plans |
Unlimited Memory vs. Chatbot Amnesia
A generic chatbot that forgets you keep leaving the horse blockable at the knee will let you hang the same piece in the next Central Cannon game. Chinese Chess Coach retains the errors that actually cost points — overextended ranked chariots, unused soldier screens, the Screen Horse answer you always repeat — plus whether you play red or black more often and what the last post-mortem concluded. That continuity is what turns a one-off explanation of Cannon Behind Horse into a pattern you stop missing.
Multi-Model AI Access
Explaining why stalemate is a loss and finding the turning point in a Same Direction Cannons middlegame are different jobs. You can switch among OpenAI, Anthropic, Google, xAI, DeepSeek, Qwen, and more so a beginner lesson stays analogical while tournament prep can use a sharper model for candidate-move comparison. ChatGPT stays on OpenAI and Claude stays on Anthropic; Xiangqi study that moves from flying-general law to endgame result tables should not be locked to one vendor.
Intelligent Tool Integration for End-to-End Workflows
Chinese Chess Coach runs a coaching loop rather than a chat: a FEN, WXF scoresheet, or spoken position goes in; material, structure, and candidates are assessed; one recommendation comes out with the principle named; a drill then targets the exact pattern you missed. Attached scoresheets and voice input keep that loop intact on a commute, not only at a desk. The product of a session is the next position you can play, not a line you cannot explain.
Xiangqi Theory That Unteaches Chess
International chess habits do not fail to transfer; they actively lose games. This agent names the false friends early — knights that no longer jump, stalemates that are losses, perpetual checks that lose for the checking side, elephants blocked at the midpoint and barred from the river — and drills screen management as the skill that decides Xiangqi, not as a footnote.
Complete beginners with no local club
Chinese Chess Coach is built for people who recognize the board but cannot read the characters on the pieces and have an evening half-hour, not a neighborhood master. Sessions start with chariot and soldier, delay the cannon until blocking is intuitive, and aim for a legal game the same day.
Chess players crossing over to Xiangqi
If Xiangqi looks like chess with a river, this agent unteaches the mappings that hang cannons and walk into flying-general pins. Screen management — creating screens, destroying them, and noticing when your own piece is a screen against you — is drilled as the skill that decides games.
Club players stuck repeating the same losses
When you stall at one strength, Chinese Chess Coach uses your own games to find cross-move patterns such as leaving a horse blockable or trading cannons while screens still exist. Follow-up drills target those patterns instead of generic puzzles.
Tournament players without a local master
Chinese Chess Coach prepares a system — Screen Horse, Single Horse, Same Direction Cannons, or Angel Pawn — as plans for which soldiers to advance, when to trade a devaluing cannon, and how to keep palace integrity. Endgame targets are chosen before middlegame trades start, including ideas drawn from the Plum Flower Manual and Secret in the Tangerine.
Players reviewing a game on the move
Paste a short game or describe the position after a river crossing and get turning points, a practical move at your level, and a drill. Speech-to-text works when you cannot type a FEN.
Free For exploring Jenova:
Plus (US$20/month) For individuals:
Premium (US$50/month) For creators & freelancers:
Pro (US$100/month) For professionals:
Max (US$200/month) For power users:
Ultra (US$500/month) For unconstrained productivity:
Enterprise (US$1,000/month) For businesses & organizations:
All plans are monthly subscriptions priced in USD, before applicable taxes. Your usage limit resets on your billing date.
Chinese Chess Coach is a specialized Xiangqi AI that trains independent whole-board judgment — screens, initiative, palace integrity, and named mates — so you can explain a position rather than copy an engine move. It infers your level from how you describe the board and coaches from first-piece questions through competitive tournament preparation. The aim is judgment you can reuse, not a single recommended move.
Chinese Chess Coach keeps your Xiangqi history — Screen Horse plans, horse-block errors, ranked-chariot overextensions — instead of dropping them at the session boundary. You can switch among models from OpenAI, Anthropic, Google, xAI, DeepSeek, Qwen, and more when a lesson needs a patient rules explainer versus a sharper Central Cannon turning-point analysis. A pasted FEN or WXF scoresheet becomes candidate moves, a named principle, and a drill on the pattern you missed, which a generic chatbot cannot run as a coaching loop.
Chinese Chess Coach takes FEN strings, WXF scoresheets, Chinese notation, attached game files, and spoken position descriptions as inputs to the same coaching loop. Google Drive, Dropbox, and Notion can hold scoresheets and repertoire notes in the conversation so analysis and the next drill share one thread. Speech-to-text covers commute reviews when you cannot type a FEN, and work syncs across web, iOS, and Android.
Yes. Chinese Chess Coach starts with the board, river, and palace, then teaches pieces from easiest to hardest so the cannon arrives after blocking already makes sense. Check, checkmate, and stalemate-as-loss come once pieces are stable; flying general and forbidden repetitions wait until you can play a legal game. Chess experience is optional.
Yes. Chinese Chess Coach accepts FEN, WXF notation such as C8.5, Chinese notation such as 炮八平五, coordinates, or a natural-language recount, and mirrors the format you used. Full-game reviews flag turning points and cross-move habits, not only the last blunder. Each recommendation includes candidate moves and the principle underneath.
Yes. Chinese Chess Coach treats chess knowledge as a set of habits to rewire rather than a shortcut. Stalemate-as-loss, perpetual check as a loss for the checking side, horses blocked at the knee, elephants that cannot cross the river, and cannons that need screens are named before they cost a winning endgame. Screen management is drilled as Xiangqi’s defining skill, not as a footnote.
Yes. Chinese Chess Coach is suitable for tournament prep that depends on plans: Screen Horse structures, cannon trades, palace integrity, and which theoretical ending to steer toward. It will also say when a line is engine-optimal but impractical at your strength. Bring a dedicated Xiangqi engine if you need raw calculation.
Yes. Chinese Chess Coach runs with full feature parity on web, iOS, and Android, including commute analysis, puzzle drills, and speech-to-text when you cannot type a FEN. Settings and memory sync, so a lesson started on your phone continues on a laptop without repeating your rating, repertoire, or last post-mortem.
Yes. Chinese Chess Coach treats Central Cannon versus Screen Horse Defense as a system with plans — which soldiers to advance, when a ranked chariot is more useful than a filed chariot, and when to trade the cannon that is about to devalue. Same Direction Cannons, Angel Pawn, Elephant Opening, and Single Horse are taught the same way. You learn why chariots develop first and why the fifth file is contested terrain, not a move tree you freeze in on move eight.
Chinese Chess Coach is a teacher; a Xiangqi engine is a calculator. An engine can rank lines and flag blunders; this agent asks what you saw, names the principle you missed, and assigns a drill at your level. Use an engine to verify a variation and this coach to understand which endgame you should have steered toward.
Yes. Chinese Chess Coach works in WXF notation for international play and in Chinese notation for commentary, manuals, and broadcast boards. Figurative explanations keep beginners moving while the characters catch up. You do not need to read Chinese to start, and you are not blocked from learning it in context.
Yes. Jenova does not use your games, FEN strings, or repertoire notes to train its own AI models. Data is encrypted in transit and at rest.
Chinese Chess Coach is part of Jenova, the AI agent platform for creativity, entertainment, and life. Jenova provides unified access to leading AI models from OpenAI, Anthropic, Google, xAI, DeepSeek, and more — alongside intelligent tool integrations for end-to-end task completion. The platform features persistent cross-session memory and runs on web, iOS, and Android, with your agents and settings synced across devices. Developed by Azeroth Inc., a New York-based technology company, Jenova is used by people around the world.
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]