#Introduction

#Dify

Dify is an open-source LLM application development platform that provides an intuitive interface for AI workflows, RAG pipelines, agent capabilities, and model management. It enables developers and non-technical users to build, deploy, and manage AI applications through a web UI.

Dify helps teams quickly build and iterate on LLM-powered applications: from simple chatbots (Assistant) and text generators to complex RAG pipelines and multi-step agent workflows. For deployment components and prerequisites (API, Worker, Web, PostgreSQL, Redis, vector store, etc.), see Install Dify.

#Core Concepts

• Application – The basic unit for organizing LLM capabilities. Types include Assistant (chat), Text Generator, Agent (reasoning and tools), and Workflow/Chatflow (visual pipelines). Each app can be published as a Web App or consumed via API.
• Workflow – Visual, node-based pipelines (LLM, retrieval, code, logic nodes) for building multi-step AI logic with drag-and-drop.
• Knowledge Base (RAG) – Datasets built from documents or URLs, with chunking, embedding, and retrieval to ground LLM responses in your data. (In the Helm chart, pgvector is the supported vector store.)
• Agent – Applications that use tools and multi-step reasoning; deploy as Web App or API.

For a concise list of main features, see Main Features. For installation and configuration on the platform, see Install Dify.

#Documentation

Dify provides comprehensive official documentation to help users understand and use the platform:

#Official Documentation

• Main Documentation: https://docs.dify.ai/
  • Core concepts, quick start, and deployment guides
  • Workflow design, RAG, and agent development
  • API reference, SDK usage, and best practices
  • Plugin development and marketplace: Installing the Plugin