home / mcp / agentify mcp server
Provides real-time task tracking and webhook notifications for AI tasks and outcomes.
Configuration
View docs{
"mcpServers": {
"gargoyle92-agentify-mcp": {
"command": "agentify-mcp",
"args": [],
"env": {
"LOG_LEVEL": "info",
"AGENTIFY_WEBHOOK_URL": "https://webhook.site/your-unique-id"
}
}
}
}You can run the Agentify MCP Server to track AI task progress and receive real-time webhook notifications for task started and task completed events. This server lets you configure a webhook endpoint and easily manage dynamic webhook URLs so your automations stay in sync with AI workflows.
You operate the MCP server by starting it locally or via a package runner, and then you trigger task lifecycle events from your AI tooling. When a task starts, call the task-started tool with a brief taskDescription. When a task finishes, call task-completed with a description of what happened and the outcome. The server will send real-time webhook notifications to the configured URL for each event, enabling you to trigger downstream actions such as logging, dashboards, or triggers in your automation stack.
Prerequisites you need installed on your computer before running the MCP server.
Install the MCP server globally so you can run the agentify-mcp command from anywhere.
# Global installation
npm install -g agentify-mcpOr run the server directly using npx without installing it globally.
# Run with npx
npx agentify-mcpConfigure the webhook URL to enable real-time notifications.
export AGENTIFY_WEBHOOK_URL="https://webhook.site/your-unique-id"
# or
export WEBHOOK_URL="https://webhook.site/your-unique-id"
agentify-mcpSet the webhook URL to enable notifications. You can provide it via environment variables or include it in your MCP configuration.
# Example environment-based MPC run
AGENTIFY_WEBHOOK_URL="https://webhook.site/your-unique-id" agentify-mcpOptionally inject webhook configuration into Claude Desktop or similar clients. You can place the webhook URL inside the client’s MCP configuration so that the client runs with the correct webhook endpoint.
{
"mcpServers": {
"agentify": {
"command": "agentify-mcp",
"env": {
"AGENTIFY_WEBHOOK_URL": "https://webhook.site/your-unique-id",
"LOG_LEVEL": "info"
}
}
}
}Basic execution without a webhook URL keeps the server running without outbound notifications.
# Run without webhook
agentify-mcpWebhook delivery may require a reachable URL and properly exported variables. If you encounter issues, verify that the environment variable is set and that the webhook URL is accessible from your network.
If you change the webhook URL, restart the MCP server to ensure the new URL takes effect.
Webhook URLs are treated as sensitive and are managed via environment variables. Enable runtime URL changes if your workflow requires rotating endpoints.
You can programmatically start and stop the MCP server from your tooling, and you can update the webhook URL at runtime if your automation requires it.
Test webhook delivery by configuring a temporary endpoint (such as a webhook catcher) and triggering a task-started or task-completed event to verify that payloads arrive as expected.
If you modify the webhook URL during a session, you can verify the new configuration by checking the webhook catcher for incoming payloads.
A simple MCP server for AI task tracking and webhook notifications, with environment-based configuration for dynamic webhook URLs.
Call when starting any task or work to begin tracking the task.
Automatically monitors long-running tasks and tracks progress.
Call when finishing any task to record the outcome and details.