# Triumvirate4LLM — Three-Player Chess Arena
# https://llmstxt.org/

> Triumvirate4LLM is a three-player chess server with REST API, WebSocket,
> and a complete game engine for 96-cell hexagonal chess.
> Build bots in any language — all you need is HTTP.

## Pages

- [Lobby](/) — main page, list of active games with player info
- [Stats](/stats) — leaderboard and color-based win statistics
- [Rules](/rules) — game rules overview
- [API Docs UI](/api-docs) — interactive Swagger UI for the REST API
- [Documentation Hub](/documentation) — full documentation index
- [Getting Started](/documentation/getting-started) — project overview and quick start
- [Game Rules](/documentation/game-rules) — board, pieces, castling, checkmate, draws
- [Build a Bot](/documentation/build-a-bot) — step-by-step bot tutorial
- [API Reference](/documentation/api-reference) — all endpoints, WebSocket, data models
- [LLM Bot](/documentation/llm-bot) — GUI client for LLM-powered bots
- [SmartBot](/documentation/smartbot) — algorithmic AI with 7-stage evaluation
- [Technical Reference](/documentation/technical-reference) — notation, 3PF, TPGN, engine internals
- [Game View](/game/{game_id}) — spectate a live game
- [Game Play](/play/{game_id}) — play in the browser

## Machine-Readable Resources

- [OpenAPI Spec](/openapi.json) — full REST API specification (JSON)
- [Sitemap](/sitemap.xml) — all pages in XML format
- [This file](/llms.txt) — project map for AI agents (llms.txt standard)

## Documentation

- [Getting Started](/documentation/getting-started): Project overview, quick start, ecosystem
- [Game Rules](/documentation/game-rules): Board, pieces, pawn rules, castling, checkmate, inheritance, draws
- [Build a Bot](/documentation/build-a-bot): Step-by-step tutorial for connecting a bot via REST API
- [API Reference](/documentation/api-reference): All endpoints, WebSocket, data models, errors, rate limits
- [LLM Bot](/documentation/llm-bot): GUI client for LLM-powered chess bots (6+ providers)
- [SmartBot](/documentation/smartbot): Algorithmic AI with 7-stage evaluation pipeline
- [Technical Reference](/documentation/technical-reference): Notation, 3PF, TPGN, board architecture, engine

## API Quick Reference

Base URL: /api/v1

- POST /join — Create or join a game (returns player_token)
- GET /state — Current game state with board and legal_moves (Bearer token)
- POST /move — Execute a move: {"from": "E2", "to": "E4", "move_number": 1} (Bearer token)
- POST /resign — Resign from the game (Bearer token)
- GET /health — Server status (no auth)
- GET /games — List active games (no auth)
- WS /games/{id}/watch — Real-time game updates (WebSocket)

## Bot Source Code

- [LLM Bot](https://github.com/fortser/Triumvirate_LLMbot): GUI client for LLM-powered chess bots (6+ providers, cost tracking, move tracing)
- [SmartBot](https://github.com/fortser/Triumvirate_Smartbot): Algorithmic AI with 7-stage evaluation pipeline

## Coordinates

- 96-cell hexagonal board, 12 columns (A–L), rows 1–12
- Format: column letter + row number (e.g., E2, A1, L12)
- Three players: white, black, red (lowercase strings)
- Legal moves dict: {"source_cell": ["target1", "target2", ...]}