Runs glif AI workflows by loading workflows, runs, and user data; supports discovery and agent loading.
Configuration
View docs{
"mcpServers": {
"glifxyz-glif-mcp-server": {
"command": "npx",
"args": [
"-y",
"@glifxyz/glif-mcp-server@latest"
],
"env": {
"GLIF_IDS": "cm2v9aiga00008vfqdiximl2m,cm2v98jk6000r11afslqvooil",
"GLIF_API_TOKEN": "YOUR_TOKEN"
}
}
}
}You can run and manage AI workflows from glif.app by hosting an MCP server locally or via npx. This server loads workflows, provides rich details about runs and users, and exposes discovery and agent loading features to power your client experience.
To use the MCP server with your client, configure one or more MCP connections that load and operate workflows from glif.app. Start with the recommended npx approach to avoid local setup, then use a local build for persistent runs. Your client will be able to run workflows, fetch workflow information, search for workflows, and, if enabled, load agents.
Prerequisites: You need Node.js installed on your machine. If Node.js is not installed, download and install it from the official Node.js website.
Option A: Run via npx (recommended) set up the MCP server directly without cloning.
Option B: Build and run from a local checkout to customize or host the server yourself.
You can connect your MCP client to the server using one of the following configurations.
{
"mcpServers": {
"glif": {
"command": "npx",
"args": ["-y", "@glifxyz/glif-mcp-server@latest"],
"env": {
"GLIF_API_TOKEN": "your-token-here"
}
}
}
}{
"mcpServers": {
"glif": {
"command": "node",
"args": ["/path/to/glif-mcp-server/build/index.js"],
"env": {
"GLIF_API_TOKEN": "your-token-here"
}
}
}
}{
"mcpServers": {
"glif": {
"command": "node",
"args": ["/path/to/glif-mcp/build/index.js"],
"env": {
"GLIF_API_TOKEN": "your-token-here",
"GLIF_IDS": "cm2v9aiga00008vfqdiximl2m,cm2v98jk6000r11afslqvooil"
}
}
}
}The server relies on an API token from your glif account. Keep your token secure and do not expose it in public configurations.
Common development tasks include installing dependencies, building, and running in development mode to watch for changes.
Run a workflow by its ID with provided inputs.
Fetch detailed information about a workflow.
Retrieve a curated list of featured workflows.
Search workflows by name or description.
List your own workflows.
Get detailed information about your user account.
List available agents with optional filtering.
Load an agent and view its details including personality and workflows.