SitefinityMCP
Enables AI clients to read and query content from a Progress Sitefinity CMS instance via its OData API. Tools are dynamically generated from the live service metadata, supporting filtering, searching, and navigation across content types.
README
Sitefinity MCP Server
A zero-dependency Model Context Protocol server that lets AI clients read and query content from a Progress Sitefinity CMS instance through its REST (OData) Web Services API.
The tool surface is generated from the live service: at startup the server
fetches the OData $metadata, parses every entity set and its fields, and bakes
the discovered type names into the tools. Point it at any Sitefinity site and it
adapts to that site's content model.
- No npm dependencies — runs on plain Node.js (built-ins only). Nothing to install.
- Two transports — Streamable HTTP (for hosting) and stdio (for local clients).
- Read-only — every tool issues
GETrequests. - Live web explorer — the same server hosts an interactive UI that exercises every tool.
Default target: https://sta.eftm2.cloud.sitefinity.com/api/default/
Live deployment:
- Web explorer: https://sitefinity-mcp-eftm2.fly.dev/
- MCP endpoint: https://sitefinity-mcp-eftm2.fly.dev/mcp
Web explorer
Open the deployment root in a browser for a single-page app (served from
public/ by the same Node process) that showcases the whole server:
- Dashboard — live capability cards generated from
tools/list, plus stats. - Assistant (optional) — ask in plain English; Claude answers by calling
the MCP tools live (enable with
ANTHROPIC_API_KEY). - Content Library — all discovered content types; click one for its full field schema, expandable relations, and a live sample of items.
- Query Builder — compose OData queries (filter/select/orderby/paging/expand).
- Search — free-text
contains()search rendered as cards. - Tool Playground — auto-generated forms for every tool, built from each tool's input schema, with the exact JSON-RPC request/response shown.
- Wire log — a drawer streaming every JSON-RPC message on the wire.
No build step — it's vanilla HTML/CSS/JS.
Tools
All tool input schemas include a type enum populated from the live service
(39 content types on the default target, including custom dynamic types like
corporatefooditems, testimonials, people).
| Tool | What it does | Sitefinity API |
|---|---|---|
sitefinity_list_content_types |
List every content type with its fields & relations | service $metadata |
sitefinity_describe_type |
Full field + navigation list for one type | $metadata |
sitefinity_query_items |
Query a collection (filter/select/orderby/paging/expand/count) | GET /{type} |
sitefinity_get_item |
Fetch one item by GUID | GET /{type}(id) |
sitefinity_search_items |
Free-text contains() search across fields |
GET /{type}?$filter= |
sitefinity_count_items |
Count items matching a filter | GET /{type}/$count |
sitefinity_get_related |
Related items via a navigation field | GET /{type}(id)/{field} |
sitefinity_get_metadata |
Entity-set names (+ raw XML) | GET /$metadata |
sitefinity_raw_get |
Arbitrary OData GET escape hatch | GET /{path} |
OData query support
$filter supports eq, ne, gt, lt, ge, le, and, or, not,
contains(), startswith(), endswith(). Sitefinity params sf_culture,
sf_site, sf_provider are exposed as the culture, site, provider args.
Configuration
All via environment variables (see .env.example).
| Variable | Default | Notes |
|---|---|---|
SITEFINITY_BASE_URL |
https://sta.eftm2.cloud.sitefinity.com |
Site origin |
SITEFINITY_SERVICE_NAME |
default |
OData service name |
SITEFINITY_AUTH_MODE |
anonymous |
anonymous or password |
SITEFINITY_USERNAME / SITEFINITY_PASSWORD |
— | Required when password |
SITEFINITY_CLIENT_ID / SITEFINITY_CLIENT_SECRET |
sitefinity / — |
OAuth client |
SITEFINITY_DEFAULT_CULTURE / SITEFINITY_SITE_ID |
— | Applied to every request |
MCP_TRANSPORT |
http |
http or stdio |
HOST / PORT |
0.0.0.0 / 8080 |
HTTP transport bind |
ANTHROPIC_API_KEY |
— | Enables the AI Assistant (server-side secret) |
ANTHROPIC_MODEL |
claude-opus-4-8 |
Model for the Assistant |
AI Assistant
The optional Assistant turns natural-language questions into live MCP tool calls
(Claude runs an agentic tool-use loop over this server's own tools). It's
off until you provide a key. The key is a server-side secret — it never
appears in the repo or reaches the browser; /api/info only exposes an
aiEnabled boolean. On Fly:
fly secrets set ANTHROPIC_API_KEY=sk-ant-... # auto-restarts the app
Auth. Public content is anonymous. For protected content set
SITEFINITY_AUTH_MODE=password; the server runs the OAuth2 password grant
against {baseUrl}/sitefinity/oauth/token and caches/refreshes the token.
Run locally
No install step — there are no dependencies.
# HTTP transport (default)
node src/index.js
# → POST JSON-RPC to http://localhost:8080/mcp
# stdio transport (for local MCP clients)
MCP_TRANSPORT=stdio node src/index.js
Smoke test the HTTP transport:
curl -s http://localhost:8080/mcp -H 'Content-Type: application/json' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'
Use with an MCP client
Hosted (HTTP)
Point any Streamable-HTTP MCP client at the deployed endpoint:
https://sitefinity-mcp-eftm2.fly.dev/mcp
Local (stdio) — e.g. Claude Desktop
{
"mcpServers": {
"sitefinity": {
"command": "node",
"args": ["/absolute/path/to/SitefinityMCP/src/index.js"],
"env": { "MCP_TRANSPORT": "stdio" }
}
}
}
Then ask: "List the Sitefinity content types", "Find news items containing 'meal', newest first", "How many newsitems are there?".
Deploy to Fly.io
fly launch --no-deploy # or use the committed fly.toml
fly deploy
The Dockerfile copies the source and runs node src/index.js —
no build or install step. Health checks hit /health.
HTTP endpoints
| Method | Path | Purpose |
|---|---|---|
| GET | / |
Web explorer UI |
| POST | /mcp |
JSON-RPC 2.0 MCP messages |
| POST | /api/chat |
AI assistant (when ANTHROPIC_API_KEY is set) |
| GET | /health |
Liveness probe |
| GET | /api/info |
Server info (JSON, incl. aiEnabled) |
Project layout
src/
config.js Environment-driven configuration
odata.js OData query-string builder + filter helpers
metadata.js Dependency-free $metadata (CSDL/XML) parser
sitefinity.js REST client (auth, requests, live introspection)
tools.js MCP tools, generated from the live schema
mcp.js JSON-RPC 2.0 protocol handler
http.js Streamable HTTP transport (+ serves the web UI + /api/chat)
static.js Path-traversal-safe static file handler
stdio.js stdio transport
anthropic.js Zero-dep Anthropic client + agentic tool-use loop
chat.js AI assistant handler (bridges MCP tools to Claude)
index.js Entry point
public/
index.html Web explorer markup
styles.css Web explorer styles
app.js Web explorer logic (vanilla ES module)
Dockerfile Container image (no install step)
fly.toml Fly.io deployment config
License
MIT
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.