About Simplifyd Cloud
Simplifyd Cloud MCP Server connects MCP-enabled clients to Simplifyd Cloud resource management. It covers workspaces, services, deployments, domains, DNS, and related platform operations through either stdio or streamable HTTP.
Key features
- 01Manage workspaces, projects, and environments
- 02Create and configure cloud services
- 03Deploy services and inspect logs
- 04Manage domains, DNS zones, and records
- 05Review metrics, members, and registry resources
Use cases
- 01Deploy or update a Simplifyd Cloud service
- 02Inspect deployment logs and service metrics
- 03Configure a domain's DNS records
- 04Manage workspace members and registry credentials
MCP support
The server exposes Simplifyd Cloud management tools to MCP clients over stdio or Streamable HTTP. Authentication is transport-specific: stdio uses a local token, while HTTP requests provide their own bearer token.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://mcp.simplifyd.comclaude mcp add --transport http com-simplifyd-cloud https://mcp.simplifyd.com{"mcpServers":{"com-simplifyd-cloud":{"type":"http","url":"https://mcp.simplifyd.com"}}}Open in Cursor{"servers":{"com-simplifyd-cloud":{"type":"http","url":"https://mcp.simplifyd.com"}}}Open in VS Code{"mcpServers":{"com-simplifyd-cloud":{"type":"http","url":"https://mcp.simplifyd.com"}}}{"mcpServers":{"com-simplifyd-cloud":{"serverUrl":"https://mcp.simplifyd.com"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
Details
From the project READMESimplifyd Cloud MCP Server
An MCP (Model Context Protocol) server for the Simplifyd Cloud platform.
It exposes tools for managing Simplifyd Cloud resources from MCP-enabled clients,
over either stdio (a local client launches the binary as a subprocess) or
streamable HTTP (a hosted, multi-user deployment). One binary serves both;
MCP_TRANSPORT selects.
Tools
| Category | Tools |
|---|---|
| Auth | login (stdio only), get-me |
| Workspaces | list-workspaces, get-workspace, create-workspace, update-workspace, get-workspace-usage, get-workspace-transactions, get-my-role, list-workspace-members, add-workspace-member, update-member-role, remove-workspace-member, fund-workspace, get-registry, list-registry-repos, list-registry-artifacts, delete-registry-tag |
| Registry credentials | list-registry-credentials, add-registry-credential, update-registry-credential, remove-registry-credential — a workspace's pull credentials for private registries (ghcr.io, Docker Hub, ...), inherited by its Docker image services by registry. A service's own credentials are set with update-service and its registry_credentials / clear_registry_credentials actions |
| Projects | list-projects, get-project, create-project, update-project |
| Environments | list-environments, get-environment, create-environment, update-environment, list-env-variables, create-env-variable, update-env-variable, delete-env-variable |
| Services | list-services, get-service, create-service, update-service, delete-service, list-private-service-access, grant-private-service-access, revoke-private-service-access, list-service-variables, add-service-variable, set-service-variables, delete-service-variable, add-service-ingress, delete-service-ingress, add-tcp-proxy, delete-tcp-proxy, set-ingress-source-ranges, add-service-config, update-service-config, delete-service-config, approve-service-changeset, discard-service-changeset, add-shared-variable |
| Deployments | list-deployments, get-deployment, deploy-service, redeploy-service, undeploy-service, get-deployment-logs |
| Metrics | get-service-metrics, get-environment-metrics — CPU, memory and disk over the last 1h, 6h or 1d as a percentage of allocation, with scaling signals |
| Tokens | list-tokens, create-token, delete-token |
| Templates | list-templates, get-template, deploy-template, get-publishable-variables |
list-email-domains, add-email-domain, verify-email-domain, remove-email-domain — create the service itself with create-service and type email | |
| Domains | search-domains, quote-domain, list-domains, get-domain, register-domain (stdio only), renew-domain (stdio only), update-domain. Buying and renewing charge the wallet only on a second, confirming call; the hosted server leaves them out, since the Claude and ChatGPT directories refuse tools that spend money, and people buy and renew in the console instead. |
| DNS | list-dns-zones, add-dns-zone, check-dns-zone, delete-dns-zone, list-dns-records, add-dns-record, update-dns-record, delete-dns-record |
Every tool carries a title and explicit readOnlyHint, destructiveHint and
openWorldHint annotations, which clients use to decide when to ask before
calling it. The classification lives in tools/annotations.go, and a test
fails if a tool is added without one.
Directory listings
server.json describes the hosted server for the
MCP Registry, which registry-aware
clients such as Lovable browse. Bump its version with each release and publish
with mcp-publisher publish.
Build
go build -o cloud-mcp .
Transports
| Variable | Default | Meaning |
|---|---|---|
MCP_TRANSPORT | stdio | stdio or http |
MCP_ADDR | :8080 | Listen address (http only) |
MCP_ALLOWED_ORIGINS | (none) | Comma-separated browser origin allowlist (http only) |
SIMPLIFYD_API_URL | https://api.cloud.simplifyd.com | API base URL |
OPENAI_APPS_CHALLENGE | (none) | Domain verification token from OpenAI's plugin portal, served at /.well-known/openai-apps-challenge (http only) |
Under HTTP the MCP endpoint answers on the bare origin and on /mcp, with
a plain /healthz for probes.
Sessions are stateless, so replicas need no sticky routing.
Authentication
Authentication is per-transport, and the two never mix:
- stdio — a single local user. The token comes from
SIMPLIFYD_API_TOKENor~/.simplifyd/config.json, and thelogintool can set it interactively. - http — many callers. Each request must carry its own
Authorization: Bearer <token>, which is used for that call and nothing else. Nothing is persisted, no credential is shared between callers, andloginis not registered, since it would write one caller's credentials to the server's disk and apply them to everyone else.
stdio
Option 1 — Environment variable (recommended for CI):
export SIMPLIFYD_API_TOKEN="your-jwt-token"
Option 2 — Interactive login:
Use the login tool to authenticate with your email and password. The token is saved to ~/.simplifyd/config.json and reused automatically on subsequent runs.
Claude Code Configuration
Add to your MCP settings (.claude/settings.json or ~/.claude/settings.json):
{
"mcpServers": {
"simplifyd-cloud": {
"command": "/path/to/cloud/mcp/cloud-mcp",
"env": {
"SIMPLIFYD_API_TOKEN": "your-jwt-token"
}
}
}
}
Or if you prefer to authenticate interactively (omit SIMPLIFYD_API_TOKEN and run login first):
{
"mcpServers": {
"simplifyd-cloud": {
"command": "/path/to/cloud/mcp/cloud-mcp"
}
}
}
http
Run the server:
MCP_TRANSPORT=http MCP_ADDR=:8080 ./cloud-mcp
Point a client at it, passing your API token as a header:
{
"mcpServers": {
"simplifyd-cloud": {
"type": "http",
"url": "https://mcp.simplifyd.com",
"headers": {
"Authorization": "Bearer your-jwt-token"
}
}
}
}
Create a token with the create-token tool, or in the dashboard.
Security
Requests carrying an Origin header are refused unless the origin is listed in
MCP_ALLOWED_ORIGINS. This blocks DNS-rebinding attacks, where a page in a
user's browser would otherwise be able to drive a hosted server. Non-browser
clients send no Origin and are unaffected.
Normal resource responses omit variable values, passwords, connection URLs, credential blobs, and config-file contents. Tools whose descriptions say they explicitly reveal credentials return sensitive data by design; use them only when needed and avoid copying their output into logs or tickets.
Deployment logs are filtered for common credential patterns, but applications can emit secrets in unexpected formats. Review log output before sharing it.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.