pokemon-world-mcp
Enables exploring a small map, battling and catching real-named Pokémon via MCP tools. Players start with a starter and collect species, with progression and evolution.
README
Pokémon World MCP
Agent playground: explore a small map, battle and catch real-named Pokémon via MCP tools.
- Public URL (planned):
https://poke.vanscoding.com - MCP endpoint:
https://poke.vanscoding.com/mcp/(local:http://127.0.0.1:8080/mcp/) - Auth: same as vans-mcp —
Authorization: Bearer vcr_sk_...(router Neonapi_keysviaDATABASE_URL) or localMCP_DEV_BYPASS_KEY - Saves / catalog: game Neon
POKEMON_DATABASE_URL(pokemon_saves,pokemon_catalog_cache); otherwise local SQLite under$HOME/.pokemon-world-mcp/ - Catalog cache: if missing or older than 24h (
CATALOG_CACHE_TTL_HOURS), fetch PokéAPI and write the game DB. Fresh cache skips the API. Long-running processes re-check viaensure_fresh. API failure uses in-memory fallback and does not write fallback to DB.
This project is separate from vans-mcp-server (Notion / Calendar / Gmail).
Goal (V1)
Collect 3 different species in your party.
Starter (required): bulbasaur | charmander | squirtle at Lv5.
Progression
- Exp (official curves): each species has a PokéAPI
growth_rateand cumulative total-exp table (six rates).party.expis total experience. - Battle yield: active Pokémon gains
floor(enemy.base_experience × enemy.level / 7)on win or catch - Level-up: when total exp reaches the next level threshold; stats scale with level; may learn moves from the species learnset
- Moves: max 4. If a new move is offered while full, resolve with
battle_action(replace_move, forget_move_name=...)orskip_learnbefore exploring/battling again - Evolution: automatic at species thresholds (e.g. bulbasaur → ivysaur at 16)
Tools
| Tool | When |
|---|---|
new_game(starter) |
Start / reset save (starter required) |
get_status |
Phase, position, party summary, pending learn |
look |
3×3 nearby tiles (exploring) |
move |
N / S / E / W — grass may encounter |
party |
Party details (level, exp, moves) |
battle_status |
In battle (or pending learn) |
battle_action |
fight + move_name, catch, run, or replace_move / skip_learn |
Tip for student agents: after each tool call, narrate the result to the human before the next action.
Local development
cd C:\Users\mz038\Desktop\peas-agent\pokemon-world-mcp
uv sync --extra dev
$env:MCP_DEV_BYPASS_KEY = "vcr_sk_dev_local_only"
# Optional — router Neon (api_keys):
# $env:DATABASE_URL = "postgresql://..."
# Optional — game Neon (saves + catalog); without this, SQLite is used:
# $env:POKEMON_DATABASE_URL = "postgresql://..."
uv run pokemon-world-mcp
Without POKEMON_DATABASE_URL, player saves and the catalog cache both live in the same SQLite file (%USERPROFILE%\.pokemon-world-mcp\pokemon_world.db, or SQLITE_PATH). Progress survives restarts.
- Health:
http://127.0.0.1:8080/health(no key) - MCP:
http://127.0.0.1:8080/mcp/with Bearer header
Production on Fly must set both DATABASE_URL (router) and POKEMON_DATABASE_URL (game) — see docs/deploy-fly.md.
Tests
uv run pytest
Cursor / student Agent config (example)
Point MCP at this server (not mcp.vanscoding.com), same course API key:
{
"mcpServers": {
"pokemon-world": {
"url": "https://poke.vanscoding.com/mcp/",
"headers": {
"Authorization": "Bearer vcr_sk_YOUR_KEY"
}
}
}
}
Local:
{
"mcpServers": {
"pokemon-world": {
"url": "http://127.0.0.1:8080/mcp/",
"headers": {
"Authorization": "Bearer vcr_sk_dev_local_only"
}
}
}
}
Map notes
- 10×10 fixed map:
path,grass,water(blocked) - ~25% encounter chance on grass steps
- Wild level ≈ party average ±2 (clamped 2–50)
- Wild species by map quadrant (no starters in the wild):
| Zone | Coordinates | Pool |
|---|---|---|
| NW | x≤4, y≤4 | pidgey, rattata, jigglypuff |
| NE | x≥5, y≤4 | pikachu, meowth, jigglypuff |
| SW | x≤4, y≥5 | geodude, meowth, rattata |
| SE | x≥5, y≥5 | gastly, geodude, pidgey |
- Blackout: return to start, party healed
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
Neon Database
MCP server for interacting with Neon Management API and databases
E2B
Using MCP to run code via e2b.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.