Jump to content

Connect SuperML | Leeroopedia MCP: Equip your AI agents with best practices, code verification, and debugging knowledge. Powered by Leeroo — building Organizational Superintelligence. Contact us at founders@leeroo.com.

Workflow:Langgenius Dify Plugin Installation and Configuration

From Leeroopedia
Knowledge Sources
Domains Plugin_Management, Integration, LLM_Ops
Last Updated 2026-02-12 07:00 GMT

Overview

End-to-end process for discovering, installing, configuring, and managing plugins on the Dify platform, from marketplace browsing through credential setup to workflow integration.

Description

This workflow covers the complete plugin lifecycle in Dify's extensible plugin ecosystem. The platform supports multiple plugin types including tools (Google Search, DALL-E, custom APIs), model providers, extensions, agent strategies, and bundles. Plugins can be installed from the Dify Marketplace, GitHub repositories, or uploaded as local packages. Each plugin may require credential configuration (API keys, OAuth flows) before use. Once configured, plugins become available as nodes in workflows, tools in agents, and capabilities across the platform.

Usage

Execute this workflow when you need to extend Dify's capabilities with third-party integrations, custom tools, additional model providers, or specialized processing capabilities. This is the workflow for adding Google Search to an agent, connecting a new LLM provider, setting up MCP (Model Context Protocol) servers, or deploying custom API tools.

Execution Steps

Step 1: Browse and Discover Plugins

Explore available plugins through the Dify Marketplace or browse installed plugins in the workspace. Filter by plugin type (tool, model, extension, agent strategy), search by keyword, or browse by category tags. Review plugin descriptions, ratings, and compatibility information.

Discovery methods:

  • Marketplace: Browse the centralized plugin repository with search and filtering
  • Installed plugins: View all plugins currently active in your workspace
  • Plugin types: Tools, models, extensions, agent strategies, bundles
  • Category and tag-based filtering
  • Version compatibility checking

Step 2: Install Plugin

Install a plugin from your chosen source. The platform supports three installation methods, each with different use cases and permission requirements.

Installation sources:

  • Marketplace: One-click installation of vetted plugins
  • GitHub: Install directly from a GitHub repository release, with version selection
  • Local upload: Upload a plugin package file for custom or private plugins

Installation process:

  • Permission validation based on workspace settings (marketplace-only, official-only, or all sources)
  • Package size validation against configured limits
  • Dependency resolution and compatibility checking
  • Plugin daemon handles installation and lifecycle

Step 3: Configure Credentials

Set up the required authentication credentials for the installed plugin. Many plugins require API keys, OAuth tokens, or other credentials to access external services. The platform provides forms for entering credentials, OAuth popup flows for third-party authorization, and secure credential storage.

Credential types:

  • API key: Direct key entry for services like OpenAI, Google, etc.
  • OAuth: Popup-based authorization flow for services like Notion, Slack
  • Custom fields: Plugin-defined credential forms with validation
  • Credentials are encrypted at rest and scoped to the workspace

Step 4: Configure Plugin Settings

Adjust plugin-specific settings beyond credentials. This includes configuring endpoint URLs, setting default parameters, managing permission levels, and customizing behavior. For MCP servers, configure the server connection and available tool registrations.

Configuration options:

  • Plugin-specific parameter forms
  • Permission management (who can use the plugin)
  • Endpoint configuration for custom deployments
  • MCP server setup and tool registration
  • Default parameter values for workflow integration

Step 5: Integrate into Applications

Use the installed and configured plugin within Dify applications. Add plugin-provided tools as nodes in workflows, enable tools in agent configurations, or use model providers for LLM inference. The plugin's capabilities become available throughout the platform once installed.

Integration points:

  • Workflow nodes: Add tool nodes that invoke plugin capabilities
  • Agent tools: Enable plugins as available tools for autonomous agents
  • Model providers: Use plugin-provided models for inference
  • Knowledge base: Use plugin-provided datasource connectors
  • Triggers: Configure plugin-based workflow triggers (webhooks, schedules)

Step 6: Manage and Update Plugins

Monitor plugin health, update to new versions, and manage the plugin lifecycle. The platform tracks available updates, handles version comparison, and supports controlled upgrade rollout.

Lifecycle management:

  • Version update notifications from marketplace
  • Controlled upgrade with version selection
  • Plugin health monitoring
  • Uninstallation with cleanup
  • Permission and access control updates

Execution Diagram

GitHub URL

Workflow Repository