nextcloud-dynamic-mcp-server
Exposes a live Nextcloud instance as an MCP server by dynamically creating tools from OpenAPI definitions of installed apps. Enables natural language interaction with Nextcloud apps like files sharing, provisioning, and calendar management.
README
Dynamic MCP Server for Nextcloud
This server exposes a live Nextcloud instance as an MCP server - flexibile for all apps installed.
Instead of shipping a fixed tool list, it queries the Nextcloud ocs_api_viewer app at startup, reads the OpenAPI descriptions for installed apps, and turns those operations into MCP tools dynamically. The result is an MCP endpoint that reflects the APIs available on the connected Nextcloud instance.
What The Server Does
- Connects to a Nextcloud instance defined by
NEXTCLOUD_URL - Uses server discovery credentials at startup and per-request credentials for tool execution
- Reads installed app APIs from
NEXTCLOUD_URL/apps/ocs_api_viewer - Creates MCP tools dynamically from the discovered OpenAPI operations
- Proxies tool calls back to the real Nextcloud REST endpoints
- Supports both
streamable-httpandstdioMCP transports
One built-in tool is always available:
nextcloud_discovery_status: returns the connected Nextcloud URL, auth mode, discovered apps, tool count, and last refresh/error state
Dynamic tools are named from the Nextcloud app id plus the OpenAPI operation id or path, for example:
files_sharing_get_shares
provisioning_api_create_user
dav_upcoming_events_get_events
Main Service Endpoints
GET /
Health and discovery endpoint. Returns:
- server name
- transport mode
- MCP path
- whether default credentials are configured
- current discovery status
Example:
curl http://localhost:8000/
/mcp
Main MCP endpoint for streamable-http clients.
Point Codex, Claude Code, or any other MCP client at:
http://localhost:8000/mcp
Requirements
- Docker and Docker Compose
- A reachable Nextcloud instance
- The Nextcloud
ocs_api_viewerapp enabled on that instance - A Nextcloud username and app token with permission to access the APIs you want to expose
Configuration
The server is configured entirely with environment variables.
| Variable | Default | Description |
|---|---|---|
NEXTCLOUD_URL |
http://nc31-app-1:80 |
Base URL of the target Nextcloud instance |
NEXTCLOUD_USERNAME |
unset | Server-side username used only for startup discovery |
NEXTCLOUD_APP_TOKEN |
unset | Server-side app token used only for startup discovery |
MCP_HOST |
0.0.0.0 |
Bind host for HTTP mode |
MCP_PORT |
8000 |
Bind port for HTTP mode |
MCP_TRANSPORT |
streamable-http |
streamable-http or stdio |
DISCOVERY_TIMEOUT_SECONDS |
30 |
Timeout for discovery and proxied requests |
LOG_LEVEL |
INFO |
Python log level |
DEBUG |
unset | Set to true to enable Starlette debug mode |
Authentication Modes
The server supports two auth patterns:
- Server-level discovery auth via
NEXTCLOUD_USERNAMEandNEXTCLOUD_APP_TOKEN - Per-request auth via MCP request headers:
X-Nextcloud-UsernameX-Nextcloud-AppToken
The server-level credentials are used only during startup discovery. Every actual tool call must provide the request headers, and the server does not fall back to the startup admin credentials for execution.
How To Start The Server
Update docker-compose.yml with your Nextcloud URL and credentials, then run:
docker compose up -d --build
The server will be available at:
http://localhost:8000/
http://localhost:8000/mcp
How Discovery Works
At startup, the server:
- Calls
GET /apps/ocs_api_viewer/appson the configured Nextcloud instance - Loads each app’s OpenAPI document from
GET /apps/ocs_api_viewer/apps/{appId} - Builds an MCP input schema from the operation parameters and request body
- Registers the operation as a callable MCP tool
Discovery uses the server's default NEXTCLOUD_USERNAME and NEXTCLOUD_APP_TOKEN.
Every tool execution uses only X-Nextcloud-Username and X-Nextcloud-AppToken. If those headers are missing, the tool call is rejected instead of falling back to the startup admin account.
If discovery fails, the server still starts and reports the error through nextcloud_discovery_status and GET /.
Client Configuration Examples
Codex
Codex can pass the Nextcloud credentials as HTTP headers:
codex mcp add nextcloud-live --url http://localhost:8000/mcp
Equivalent ~/.codex/config.toml example:
[mcp_servers.nextcloud-live]
url = "http://localhost:8000/mcp"
http_headers = { X-Nextcloud-Username = "NEXTCLOUD_USERNAME", X-Nextcloud-AppToken = "NEXTCLOUD_APP_TOKEN" }
Claude Code
CLI example:
claude mcp add --transport http nextcloud-live http://localhost:8000/mcp
Project-scoped .mcp.json example using per-user credentials from environment variables:
{
"mcpServers": {
"nextcloud-live": {
"type": "http",
"url": "http://localhost:8000/mcp",
"headers": {
"X-Nextcloud-Username": "${NEXTCLOUD_USERNAME}",
"X-Nextcloud-AppToken": "${NEXTCLOUD_APP_TOKEN}"
}
}
}
}
This is useful when you want one shared MCP server URL but each developer should authenticate to Nextcloud with their own account.
Example Smoke Checks
Check the HTTP health endpoint:
curl http://localhost:8000/
If you are using Docker Compose:
docker compose logs -f mcp
In an MCP client, call:
nextcloud_discovery_status
Then verify that the discovered tool list includes operations from your enabled Nextcloud apps.
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.