Dify

Dify is a source-available platform for building AI applications through a visual workspace. Teams can assemble chatbots, agents and multi-step workflows; connect model providers and tools; ingest documents into knowledge bases; test prompts; and publish the result as a hosted web app or call it through REST APIs. Dify is available as a managed cloud service and as a self-hosted Community Edition deployed with Docker Compose.

The repository uses the Dify Open Source License, which is based on Apache License 2.0 but adds material conditions. Operating the source code as a multi-tenant service requires written authorization or a commercial license, and Dify's logo and copyright notices cannot be removed or modified in the covered frontend. The terms are not equivalent to unmodified Apache-2.0. Organizations planning a hosted service, white-label deployment or redistribution should review the current license text rather than assume standard Apache permissions.

Top features

• Visual workflow builder: Connect prompts, model calls, tools, retrieval, code and control-flow steps on a canvas, then test the path with visible inputs and outputs.
• Agent applications: Configure agents that use function calling or ReAct-style reasoning and give them built-in, marketplace or custom tools.
• Knowledge and RAG pipelines: Import common document formats, process content into a knowledge base and retrieve relevant passages inside chat or workflow applications.
• Model management: Connect proprietary, open-source, self-hosted and OpenAI-compatible model endpoints from the same workspace.
• Publishing and APIs: Release a Dify project as a web application or integrate it into another product through generated backend APIs.
• Plugins and marketplace: Extend models, tools and integrations with plugins, and install community-provided components and templates.
• Cloud or self-hosting: Use Dify Cloud for managed operation or run the Community Edition with the documented Docker Compose stack.

Use cases

• Build an internal knowledge assistant over policies, product documentation or support material.
• Create a tool-using customer-service or operations agent with a reviewable visual execution path.
• Turn a repeatable content, research or data-processing procedure into an API-backed workflow.
• Self-host an AI application workspace when deployment control is required, subject to Dify's additional license conditions.