depot-mcp
Gives LLMs structured access to Warhammer 40,000 (10th edition) game data: faction lists, unit datasheets, stratagems, detachments and enhancements.
README
depot-mcp
An MCP server that gives LLMs structured access to Warhammer 40,000 (10th edition) game data: faction lists, unit datasheets, stratagems, detachments and enhancements. The data comes from a depot instance, which in turn builds it from Wahapedia's CSV exports.
Models answering rules questions from memory get stats wrong constantly. Giving them lookup tools instead of trivia recall fixes that: the model calls get_datasheet("tyranids", "Barbgaunts") and quotes the actual sheet.
Tools
| Tool | What it does |
|---|---|
list_factions |
Every faction with its slug, datasheet count, detachment count and data version. |
search_datasheets(query, faction?) |
Find units by name substring, optionally within one faction. |
get_datasheet(faction, datasheet) |
Full unit rules: model stats, abilities, wargear profiles, composition, points, keywords, options, leader attachments. |
search_stratagems(query?, faction?) |
Search core and detachment stratagems by name, type or phase. |
get_detachment(faction, detachment?) |
Detachment abilities, enhancements and stratagems. Omit detachment to list a faction's detachments. |
All tools are read-only. HTML in the source data is flattened to plain text before it reaches the model.
Getting the data
The server reads depot's generated JSON (index.json, per-faction faction.json, per-unit datasheet files). Point it at either:
- A running depot instance (
DEPOT_DATA_URL): any depot deployment serves its data under/data/. If you self-host depot, use that URL. - A local directory (
DEPOT_DATA_DIR): generate the JSON yourself with depot's CLI, no web app needed:
git clone https://github.com/fjlaubscher/depot
cd depot
pnpm install
pnpm --filter @depot/core build && pnpm --filter @depot/cli build
pnpm --filter @depot/cli start # downloads Wahapedia CSVs, writes JSON
# data is now in packages/cli/dist/data
Set exactly one of the two variables. Data is loaded once at startup; restart the server to pick up refreshed data.
Running
With uv:
DEPOT_DATA_DIR=/path/to/depot/packages/cli/dist/data uv run depot-mcp
With Docker:
docker build -t depot-mcp .
docker run -p 11437:11437 \
-e MCP_HOST=0.0.0.0 \
-e DEPOT_DATA_URL=https://your-depot-instance.example \
depot-mcp
Configuration
| Variable | Default | Purpose |
|---|---|---|
DEPOT_DATA_URL |
Base URL of a depot instance. Set this or DEPOT_DATA_DIR, not both. |
|
DEPOT_DATA_DIR |
Path to depot CLI output (the directory containing index.json). |
|
MCP_HOST |
127.0.0.1 |
Bind address. Use 0.0.0.0 in containers. |
MCP_PORT |
11437 |
Listen port. |
MCP_AUTH_TOKEN |
empty | Static bearer token. Empty disables auth; only do that on a trusted network. |
LOG_LEVEL |
INFO |
Python log level. |
Transport is streamable HTTP at /mcp. Don't expose the server to the internet directly; keep it on a LAN or behind a VPN.
Connecting clients
Claude Code:
claude mcp add depot \
--transport http \
--url http://localhost:11437/mcp \
--header "Authorization: Bearer $MCP_AUTH_TOKEN"
Open WebUI: Admin Settings → External Tools → add an MCP connection with URL http://<host>:11437/mcp, auth type Bearer, and your token.
Any other MCP client: standard streamable-http endpoint, JSON-RPC over POST:
curl -X POST http://localhost:11437/mcp \
-H "Authorization: Bearer $MCP_AUTH_TOKEN" \
-H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'
Development
uv sync
uv run pytest
uv run mypy
Tests run against a committed subset of real depot output in tests/data, so they need no network and no running depot.
License
MIT. This is a fan project: not affiliated with, endorsed by, or sponsored by Games Workshop Limited. Game data comes from Wahapedia's public exports via depot; the underlying game rules are Games Workshop's intellectual property.
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.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
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.