Dify
Open-source LLM app development platform that combines a visual AI workflow canvas, RAG pipelines, agent tooling and model management in one workspace, for teams shipping production-ready AI apps.
Overview
Dify is an open-source platform for building LLM applications. It combines a visual AI workflow canvas, a RAG pipeline, agent capabilities (function calling or ReAct, with 50+ built-in tools), a Prompt IDE, model management across hundreds of providers, and LLMOps observability (via Opik, Langfuse or Arize Phoenix) in one workspace.
How it works
- Design a multi-step AI workflow on the visual canvas (prompt chains, conditionals, tool calls, retrieval steps).
- Wire in a RAG pipeline: ingest PDFs/PPTs/docs, chunk and index them, and retrieve context at query time.
- Attach agent capabilities so the app can call tools (Google Search, DALL·E, Stable Diffusion, WolframAlpha, or your own) using function calling or ReAct.
- Every app is automatically backed by an API, so it can be embedded in your own product (Backend-as-a-Service).
- Monitor production traffic and iterate on prompts/datasets using the built-in LLMOps dashboard.
Licensing note
Dify is not plain Apache-2.0: it ships under the Dify Open Source
License, an Apache-2.0 base with two extra conditions. First, you may not
operate a multi-tenant SaaS using the source code without a commercial
license from LangGenius (one tenant equals one workspace). Second, you may
not remove or modify the LOGO or copyright notices in the frontend (web/).
Self-hosting Dify for your own business, including commercial use, is
explicitly allowed; enterprise features have their own separate commercial
license.
Examples
- A RAG chatbot over internal company documents with citation and observability built in.
- A multi-step customer-support agent that calls internal APIs as tools and escalates to a human when confidence is low.
- Prototyping a prompt in the Prompt IDE, then promoting the same app to production via the auto-generated API.
Installation
You need Docker (with the Compose plugin) and Git. Dify is self-hosted, so there's no managed cloud account needed to try it.
- Clone the repo and move into the Docker setup:
git clone https://github.com/langgenius/dify.git --depth 1 cd dify/docker - Copy the example environment file:
cp .env.example .env - Start the stack:
docker compose up -d - Open
http://localhost/installin your browser and finish the setup wizard to create the admin account. - In Settings, add a model provider (OpenAI, Anthropic, a local Ollama instance, or another supported provider) before you build your first app.
Keep the licensing note above in mind if you're planning to run Dify as a multi-tenant SaaS.
Related assets
Langflow
Visual Python platform for building and deploying AI agents and workflows, with full component source access, multi-agent orchestration and one-click export as an API or MCP server.
Graphify
Turns a codebase, plus its docs, PDFs, images, and video, into a queryable knowledge graph that a coding agent can traverse instead of grepping raw files, and never guesses silently: every edge is tagged EXTRACTED, INFERRED, or AMBIGUOUS so you know what was found versus inferred.
Context7
Pulls updated, version-specific documentation and code examples straight from the source into your prompt, so agents stop citing outdated APIs or hallucinating ones that don't exist.