MCP Server
Gumlet MCP server generated from the Gumlet OpenAPI spec. Connect it to Cursor, Claude Code, Windsurf, Claude Desktop, or any MCP client so the model can call Gumlet API operations without you pasting request examples into chat.
Server URL: https://mcp.gumlet.com/mcp/v1
API base: https://api.gumlet.com/v1
Listing tools works without a key. Live calls need a Gumlet API key.
What the Gumlet MCP provides
| Capability | What it does |
|---|---|
| List tools | Discover available Gumlet API operations |
| API tools | Call Gumlet REST operations from the OpenAPI document |
Connect your client
Most clients should use the hosted server.
Cursor
Add this to ~/.cursor/mcp.json (global) or .cursor/mcp.json (this project):
{
"mcpServers": {
"gumlet": {
"url": "https://mcp.gumlet.com/mcp/v1",
"headers": {
"GUMLET_API_KEY": "YOUR_GUMLET_API_KEY"
}
}
}
}
Reload the window, then ask the agent something like “List my Gumlet workspaces.”
Claude Code
claude mcp add --transport http gumlet https://mcp.gumlet.com/mcp/v1 \
--header "GUMLET_API_KEY: YOUR_GUMLET_API_KEY"
Windsurf
Add this to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"gumlet": {
"url": "https://mcp.gumlet.com/mcp/v1",
"headers": {
"GUMLET_API_KEY": "YOUR_GUMLET_API_KEY"
}
}
}
}
Claude Desktop
Add this to claude_desktop_config.json:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"gumlet": {
"url": "https://mcp.gumlet.com/mcp/v1",
"headers": {
"GUMLET_API_KEY": "YOUR_GUMLET_API_KEY"
}
}
}
}
If your Claude Desktop build does not support remote HTTP MCP yet, use the local stdio config below, or mcp-remote:
{
"mcpServers": {
"gumlet": {
"command": "npx",
"args": [
"mcp-remote",
"https://mcp.gumlet.com/mcp/v1",
"--header",
"GUMLET_API_KEY: ${GUMLET_API_KEY}"
],
"env": {
"GUMLET_API_KEY": "YOUR_GUMLET_API_KEY"
}
}
}
}
Other MCP clients
Point any MCP-compatible client at:
https://mcp.gumlet.com/mcp/v1
Send your key in the GUMLET_API_KEY header.
Try it
After the server is connected, start a chat and ask:
- List my Gumlet workspaces.
- Create a video asset in this collection.
- Show analytics for this video.
- Create a TypeScript example that creates a video asset.
The assistant should call Gumlet API tools through MCP instead of guessing request shapes from training data.
Ask AI on the docs site
Ask AI on these pages searches published guides in the browser. Use MCP when you want the model to run Gumlet API operations from Cursor, Claude, or another agent.
