Multi-Domain MCP for Spiritual Intelligence: 9+ Servers, 120+ Tools, One API Key
Connect 9+ spiritual data domains to any AI agent with one API key. MCP servers for astrology, tarot, numerology, I-Ching, dreams, crystals, and more.
TL;DR
- RoxyAPI ships 9+ dedicated MCP servers covering astrology, Vedic astrology, tarot, numerology, I-Ching, dreams, crystals, angel numbers, and location data, all under one API key.
- Each server exposes domain-specific tools that AI agents auto-discover. Over 120+ tools total, zero manual tool definitions needed.
- Works with Claude Desktop, Cursor, Windsurf, VS Code Copilot, Claude Code, and any MCP-compatible client.
- Remote HTTP transport means no npm install, no local process, no dependency management. Add a URL and start building.
About the author: Brett Calloway is a Developer Advocate and AI Integration Specialist with 12 years of experience building APIs and developer tooling, including three years focused on AI-native infrastructure for spiritual and wellness applications. He has led developer relations at two Series B SaaS companies and spoken at PyCon and JSConf on building context-rich AI agents using Model Context Protocol.
Model Context Protocol (MCP) gives AI agents a standardized way to discover and call external tools. Instead of hardcoding function definitions, the agent asks the MCP server "what tools do you have?" and gets back a typed schema it can call on demand. This is a massive improvement over manual tool wiring. But most MCP servers in the spiritual and wellness space cover a single domain. One server for astrology. Another for tarot. Maybe a third for numerology. Each requires its own API key, its own configuration, and its own billing. If you want your AI agent to pull a birth chart, draw a tarot card, and interpret a dream symbol in the same conversation, you are stitching together three separate providers with three separate auth flows. That is the problem multi-domain MCP solves. This article shows how a single API key connects 9+ spiritual data domains to any MCP-compatible AI agent, with working configuration examples for Claude Desktop, Cursor, and Claude Code.
What Multi-Domain MCP Looks Like in Practice
What Multi-Domain MCP Looks Like in Practice
Roxy ships a dedicated MCP server for every data domain in the platform. Each server has its own endpoint, its own set of tools, and its own OpenAPI-derived schema. But they all share one API key and one subscription.
| Domain | MCP Endpoint | Example Tools |
|---|---|---|
| Western Astrology | /mcp/astrology-api | Daily horoscope, birth chart, synastry, transits |
| Vedic Astrology | /mcp/vedic-astrology-api | Kundali, nakshatras, dasha periods, KP analysis |
| Tarot | /mcp/tarot-api | Daily card, spreads, card lookup by ID |
| Numerology | /mcp/numerology-api | Life path, expression, compatibility, personal year |
| I-Ching | /mcp/iching-api | Hexagram reading, changing lines, trigram lookup |
| Dreams | /mcp/dreams-api | Symbol search, category browse, interpretation |
| Crystals | /mcp/crystals-api | Crystal properties, chakra associations, zodiac stones |
| Angel Numbers | /mcp/angel-numbers-api | Number meaning, sequence interpretation |
| Location | /mcp/location-api | Timezone, coordinates, city search |
One key. One subscription. 9+ servers. 120+ tools. The agent discovers what it needs and calls it. You do not wire anything manually.
Ready to connect your AI agent to 9+ spiritual domains? See all products and get your API key in under two minutes. View pricing.
How MCP Server Discovery Works for AI Agents
When you register a Roxy MCP server with your AI client, the agent sends a tools/list request. The server responds with every available tool, including typed parameter schemas, descriptions, enum values, and examples. The agent uses these descriptions to decide when and how to call each tool.
This is the key difference between MCP and manual function calling. With function calling, you write JSON schema for each tool yourself and embed it in your system prompt. With MCP, the server provides the schema automatically. Add a new endpoint to the API, and the agent discovers it on the next tools/list call. No prompt updates, no schema maintenance.
Each tool description is written for three audiences: human developers reading docs, AI agents choosing which tool to call, and search crawlers indexing the API surface. Rich descriptions with domain keywords (nakshatra, hexagram, Life Path, Major Arcana) help agents make better tool selection decisions.
Setting Up Multi-Domain MCP in Claude Desktop and Cursor
Each product domain has its own remote MCP server at https://roxyapi.com/mcp/{product-slug}. Register the domains your agent needs. Visit the MCP Setup page for the exact config for your platform.
Claude Desktop (via claude_desktop_config.json):
{
"mcpServers": {
"roxy-astrology-api": {
"type": "http",
"url": "https://roxyapi.com/mcp/astrology-api",
"headers": {
"X-API-Key": "YOUR_API_KEY"
}
},
"roxy-tarot-api": {
"type": "http",
"url": "https://roxyapi.com/mcp/tarot-api",
"headers": {
"X-API-Key": "YOUR_API_KEY"
}
}
}
}
Cursor (via .cursor/mcp.json in your project root):
{
"mcpServers": {
"roxy-astrology-api": {
"url": "https://roxyapi.com/mcp/astrology-api",
"headers": {
"X-API-Key": "${env:ROXY_API_KEY}"
}
},
"roxy-tarot-api": {
"url": "https://roxyapi.com/mcp/tarot-api",
"headers": {
"X-API-Key": "${env:ROXY_API_KEY}"
}
}
}
}
Claude Code (one command per domain):
claude mcp add --transport http roxy-astrology-api https://roxyapi.com/mcp/astrology-api --header "X-API-Key: YOUR_KEY"
claude mcp add --transport http roxy-tarot-api https://roxyapi.com/mcp/tarot-api --header "X-API-Key: YOUR_KEY"
No npm install. No local process to manage. No dependency conflicts. The server runs remotely over Streamable HTTP, and the client discovers tools on connection. Add as many domains as your agent needs.
Choosing Which Domains to Register
Register the domains your agent needs. Each product has its own MCP URL at https://roxyapi.com/mcp/{product-slug}.
For example, an astrology chatbot might only need Western and Vedic astrology:
{
"mcpServers": {
"astrology": {
"type": "http",
"url": "https://roxyapi.com/mcp/astrology-api",
"headers": {
"X-API-Key": "YOUR_API_KEY"
}
},
"vedic-astrology": {
"type": "http",
"url": "https://roxyapi.com/mcp/vedic-astrology-api",
"headers": {
"X-API-Key": "YOUR_API_KEY"
}
}
}
}
Same API key for both. The agent sees only the tools relevant to its task, which means fewer tool descriptions in context and more precise tool selection. This pattern works well when you want tight control over what the agent can access.
A Multi-Domain Conversation: What the Agent Actually Does
Here is what happens when a user asks their AI agent a question that spans multiple spiritual domains. Imagine the prompt: "Tell me about my day. I am an Aries born on March 28, 1990."
The agent, with access to all Roxy MCP servers, can chain tool calls across domains in a single response:
- Astrology: calls
get_horoscope_aries_dailyto pull the daily horoscope for Aries - Tarot: calls
get_tarot_dailyto draw a daily guidance card - Numerology: calls
get_personal_daywith the birth date and current date to calculate the personal day number - Angel Numbers: calls
get_angel_number_328if the date matches a significant sequence
The agent weaves these results into a coherent narrative: the horoscope sets the tone, the tarot card adds symbolic guidance, the personal day number provides numerological context, and the angel number ties it together. Four domains, four tool calls, one natural response. No custom orchestration code. The agent handles the routing.
This kind of cross-domain synthesis is impossible with single-domain MCP servers unless you manually configure and authenticate four separate providers. With Roxy, the agent handles all routing and domain selection using the tool descriptions provided by each MCP server. You build the user experience. The infrastructure handles the rest.
Remote HTTP MCP vs. Local stdio Servers
Most MCP servers in the ecosystem ship as npm packages that run locally via stdio transport. You install them with npx, they spawn a child process, and your AI client communicates over stdin/stdout. This works, but it creates operational overhead: dependency installation, version management, process supervision, and platform-specific binary issues.
Roxy uses remote Streamable HTTP transport exclusively. The MCP server runs on Roxy infrastructure. Your client sends a POST request, gets a JSON-RPC response. No local process. No node_modules. No build step.
| Feature | Remote HTTP (Roxy) | Local stdio |
|---|---|---|
| Setup | Add URL + API key | npm install + spawn process |
| Updates | Automatic (server-side) | Manual npm update |
| Dependencies | None | Node.js runtime required |
| Platform support | Any HTTP client | OS-specific binary issues |
| Scalability | Server-managed | Local resource limits |
| Auth | Header-based, standard | Environment variable wiring |
The tools/list endpoint is public and free. Only tools/call (the actual data request) counts as a billable API call, identical to a REST request. This means agents can discover and inspect tools without consuming quota.
Compatible AI Clients and Frameworks
Multi-domain MCP works with any client that speaks the Model Context Protocol. The current ecosystem includes:
- Claude Desktop: native MCP support via config file
- Cursor: MCP via settings UI or
.cursor/mcp.json - Claude Code: register via CLI command
- Windsurf: MCP server configuration in settings
- VS Code Copilot: MCP support via extensions
- OpenAI Agents SDK: MCP client integration
- LangChain / LangGraph: MCP tool provider adapters
- CrewAI: MCP as tool source for multi-agent workflows
- Custom agents: any implementation using the MCP TypeScript or Python SDK
The protocol is standardized. If your client can send a JSON-RPC POST request with an API key header, it can use Roxy MCP servers. The full list of supported tools and their schemas is available at the API reference, where you can test endpoints live before wiring them into your agent.
Frequently Asked Questions
Q: Can one MCP server cover multiple spiritual domains?
A: Yes. Register multiple per-domain MCP servers (e.g., https://roxyapi.com/mcp/astrology-api, https://roxyapi.com/mcp/tarot-api) under the same API key. Your agent discovers tools from all registered domains. See the MCP Setup page for all product URLs.
Q: How many tools can an AI agent realistically use in one conversation? A: Modern AI models handle dozens of tools in their context window. Roxy exposes 120+ tools across all domains, and agents select the relevant ones based on tool descriptions. In practice, a single conversation typically calls 3 to 8 tools across 2 to 4 domains. The agent manages tool selection automatically.
Q: What is the difference between remote HTTP MCP and local stdio MCP servers? A: Remote HTTP servers run on the provider infrastructure. You add a URL and authenticate with a header. Local stdio servers run as a child process on your machine, requiring npm install and a Node.js runtime. Remote HTTP eliminates dependency management and receives automatic updates server-side.
Q: Which AI coding tools support MCP for spiritual data? A: Claude Desktop, Cursor, Claude Code, Windsurf, VS Code Copilot, and OpenAI Agents SDK all support MCP natively. LangChain, CrewAI, and any client using the MCP SDK can also connect. See the MCP setup guide for configuration instructions for each platform.
One Key, Nine Domains, Zero Configuration Overhead
Multi-domain MCP is not a feature. It is an architecture decision. When your AI agent can access birth charts, tarot readings, dream interpretation, I-Ching hexagrams, numerology profiles, crystal properties, and angel number meanings through a single protocol and a single API key, you stop building plumbing and start building experiences.
Roxy is the only MCP server covering 9+ spiritual and insight domains under one key. No credential sprawl. No multi-vendor billing. No tool schema maintenance.
Explore the full API reference to see every endpoint and tool available. Get your API key and connect your first MCP server in under five minutes.