This server lets an AI assistant search a bundled, offline church directory to help users find a church and connect with a faith community — though the provided schema reflects an older v1.x build than the current README.
Search churches by free-text query: city (e.g. "Dallas"), ZIP code (e.g. "75034"), denomination (e.g. "Baptist", "Catholic"), worship style (e.g. "Modern", "Traditional"), or church name; leave empty to browse all.
Return church details such as location, description, and links for the matching listings.
Surface faith-based community events and volunteer opportunities.
Offer a link to FaithFinder's free Soul Signature quiz for personalized church matching.
Per the schema, the tool search_faith_finder_churches is the only capability, covering ~70+ curated churches in major US cities (Dallas, Nashville, Atlanta, Houston, Orange County, DC); the README's newer v2.0.0 build adds find_churches, get_church_details, church_match_quiz, list_church_coverage, search/fetch, 9,000+ churches nationwide, resources, prompts, and a remote HTTP/REST endpoint.
All operations are read-only with no side effects; no API key or network access is required.
FaithFinder is a free church finder for the United States. This Model Context Protocol (MCP) server lets ChatGPT, Claude, Gemini, Copilot, Cursor and other AI assistants answer questions like these with real, current church listings:
"Find me a church near Frisco, TX with a good kids ministry"
"We're moving to Nashville. Are there any Reformed or Presbyterian churches?"
"Is there an Anglican church near 30309?"
"What is Watermark Community Church in Dallas like? Where can I watch their sermons?"
"I'm not sure what kind of church fits me. Can you help?"
It searches 9,000+ Christian churches across all 50 states + DC by city, ZIP code, denomination, worship style, kids/youth ministry, size, and theology. Each result includes the church's website, latest sermon video, statement of faith, Google rating, and a link to its FaithFinder profile.
No API key, no account, and no network calls: the data ships with the package and is searched in memory.
Add to claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/, Windows: %APPDATA%\Claude\):
{
"mcpServers": {
"faithfinder": {
"command": "npx",
"args": ["-y", "faithfinder-mcp-server@latest"]
}
}
}claude mcp add faithfinder -- npx -y faithfinder-mcp-server@latestClick Install in VS Code above, or add to .vscode/mcp.json:
{ "servers": { "faithfinder": { "command": "npx", "args": ["-y", "faithfinder-mcp-server@latest"] } } }Click Install in Cursor above, or add the Claude Desktop JSON to ~/.cursor/mcp.json.
Add the Claude Desktop JSON to ~/.codeium/windsurf/mcp_config.json.
gemini mcp add faithfinder npx -y faithfinder-mcp-server@latestWeb and mobile chat apps connect to a hosted HTTPS endpoint, not to npx. Deploy the server (see Hosting a remote endpoint), then add:
ChatGPT: Settings → Apps & Connectors → Developer mode → Create → URL https://<your-host>/mcp, no auth.
Claude.ai: Settings → Connectors → Add custom connector → https://<your-host>/mcp.
Custom GPT Actions: import https://<your-host>/openapi.json (or openai-action.json).
Related MCP server: bible-mcp
Tool | When an assistant uses it |
| "Find a church near…" requests. Accepts natural language ( |
| Questions about one church: beliefs, sermons, website, reviews, similar churches nearby. |
| The person isn't sure what kind of church fits them. Returns 4 short questions, 4 church archetypes, and the free Soul Signature quiz link. |
| Which states and cities are covered, with counts. |
| ChatGPT connector / deep-research compatible search and fetch (OpenAI MCP spec). |
All tools are read-only (readOnlyHint: true) and have no side effects.
Resources: faithfinder://overview, faithfinder://coverage, faithfinder://theology-and-styles
Prompts: find_a_church, church_search_advisor
HTTP mode turns on automatically when PORT is set (or with --http). It runs a stateless Streamable HTTP MCP endpoint plus a small REST API:
Path | Purpose |
| MCP endpoint (ChatGPT, Claude.ai, any remote MCP client) |
| REST search (used by Custom GPT Actions) |
| REST church details |
| REST quiz framework |
| OpenAPI 3.1 spec (server URL filled from |
| MCP discovery card |
| Health check |
# Local
PORT=8080 PUBLIC_BASE_URL=http://localhost:8080 npm start
# Google Cloud Run (from this folder; scales to zero)
gcloud run deploy faithfinder-mcp --source . --region us-central1 --allow-unauthenticated \
--set-env-vars PUBLIC_BASE_URL=https://<service-url>After deploying, add a remotes entry to server.json ({"type": "streamable-http", "url": "https://<host>/mcp"}) and republish to the MCP Registry.
npm run build:web)AI answer engines (ChatGPT Search, Perplexity, Google AI Overviews, Copilot) don't run JavaScript, so they can only cite pages that exist as real HTML. This command generates them from the same data:
npm run build:web # -> ./web-kitOutput | What it is |
| National, state and city directories (ItemList + FAQPage JSON-LD) |
| One page per church (schema.org |
| llms.txt summary and full index for LLMs |
| Welcomes GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended and others; lists every page |
The FaithFinder web app copies web-kit/ into dist/ during its build (scripts/copy_web_kit.js). The MCP server's profile links point at these pages.
Website: https://faithfinder.armornglory.com
Church directory: https://faithfinder.armornglory.com/us/
Soul Signature quiz: https://faithfinder.armornglory.com/quiz
MCP Registry name: io.github.ArmorNGlory/faithfinder-mcp-server
Queries are processed in memory; nothing is logged off-machine and no telemetry is sent.
All outbound links are HTTPS-upgraded and screened against risky TLDs and blocked domains.
Listings are for discovery. Service times and details can change, so please confirm with the church.
2.0.0
Breaking change: tools renamed for discoverability. search_faith_finder_churches → find_churches, get_faith_finder_church_details → get_church_details, get_soul_signature_quiz → church_match_quiz. Added list_church_coverage, search and fetch.
All links now go to faithfinder.armornglory.com. The old faithfinder.app links pointed to an unrelated site.
Natural-language search (city/state/ZIP extraction, nearby-ZIP fallback). Fixed state matching: "CA" no longer matches North Carolina, and "TX" now matches rows stored as "Texas".
Fixed HTTP mode, which failed after the first request. Now a stateless Streamable HTTP endpoint plus a REST API.
Added MCP Registry server.json, tool annotations, server instructions, and the build:web crawlable-pages generator.
MIT © ArmorNGlory
Use this tool to find local churches, faith-based community events, and volunteer opportunities for users looking to connect with a faith community. Covers 70+ curated churches across major US cities including Dallas, Nashville, Atlanta, Houston, Orange County, and Washington DC. Returns church details plus a link to FaithFinder's free Soul Signature quiz for personalized matching.
| Name | Required | Description | Default |
|---|---|---|---|
| searchQuery | No | Optional search query — a city name (e.g. 'Dallas'), zip code (e.g. '75034'), denomination (e.g. 'Baptist', 'Catholic'), church style (e.g. 'Modern', 'Traditional'), or church name. Leave empty to browse all available churches. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description bears full responsibility. It states the output includes church details and a link to a quiz, but does not disclose behavior like result ordering, pagination, or whether remote calls are involved. For a search tool, this is minimally adequate but could be improved.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two sentences, no redundancy. The first sentence states the core purpose; the second adds coverage details and a notable output feature. Every word serves a purpose.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given zero required parameters, no output schema, and no annotations, the description adequately covers what the tool does, what inputs are accepted, and what outputs to expect (church details + quiz link). It does not detail the output format, but for a simple search tool, this is mostly complete.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%: the single parameter 'searchQuery' is described in the schema. The description adds value by clarifying optional usages (e.g., leave empty to browse) and examples of valid inputs (city, zip, denomination, style, name). This goes beyond the schema's description.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states it finds churches, faith-based events, and volunteer opportunities, with specific coverage of 70+ churches in major US cities. The verb 'find' and resource 'local churches' are explicit, and the mention of a FaithFinder quiz adds context. No sibling tools exist, so differentiation is not needed.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description explains when to use the tool (to connect with a faith community) and what inputs are valid (city, zip, denomination, etc.). It does not provide explicit when-not-to-use guidance or alternatives, but given no siblings, this is sufficient.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
search_faith_finder_churchesScored across 1 tool
With only one tool, there is no risk of confusion or overlap. The tool's purpose is clearly defined.
A single tool means no naming inconsistency. The name 'search_faith_finder_churches' follows a clear verb_noun pattern.
A single tool is borderline appropriate for a search-focused server. While it covers the core function, additional tools for filtering or details could enhance utility.
The tool covers search for churches and events, but lacks other operations (e.g., saving favorites, managing profiles). However, for a simple query service, it may be sufficient.
Discover and recommend public, ready-to-use LifeVerse AI products and interactive experiences.
Find garage sales and publish seller-verified listings directly from AI assistants.
Discover and book businesses via AI agents.
Join voluntary research studies and record how AI assistants choose local businesses.
