# Dify

> Open-source LLM application development platform with visual orchestration, RAG, and enterprise LLMOps

Dify is a comprehensive open-source platform for developing, deploying, and managing Large Language Model (LLM) applications. Launched in March 2023 under the Apache License 2.0, it has quickly established itself as a powerful solution for building sophisticated AI workflows with minimal coding.

At its core, Dify provides a visual orchestration studio that allows developers and teams to design complex AI applications through an intuitive drag-and-drop interface. This visual approach to workflow design has been widely praised for making prompt engineering and application architecture accessible to both technical and non-technical users. The platform supports multiple orchestration modes, including simple orchestration for basic applications, assistant orchestration for more interactive experiences, and flow orchestration for complex, multi-step workflows.

Dify excels in Retrieval-Augmented Generation (RAG) capabilities with an industry-first visual knowledge base management interface. It supports various indexing methods (keywords, text vectors, LLM-assisted question-snippet models) and retrieval techniques (keywords, text similarity matching, hybrid search, multi-path retrieval). The platform can process multiple document formats (TXT, Markdown, PDF, HTML, DOC, CSV) and integrates with numerous vector databases including Qdrant, Weaviate, Milvus, Pgvector, and more.

For developers building more advanced applications, Dify includes robust agentic workflow features with live-editing node debugging, modular DSL, and native code runtime. The platform supports a wide range of LLMs from providers like OpenAI, Anthropic, and others, while also allowing integration with local model inference runtimes like Xorbits, OpenLLM, and Ollama. Additionally, it provides tooling support for OpenAI Plugin standard tools and can directly load OpenAPI Specification APIs.

Dify's enterprise features include comprehensive logging, content moderation, team collaboration through workspaces, and secure deployment options via Docker and Helm. With weekly releases and a codebase exceeding 130,000 lines, Dify continues to evolve rapidly to meet the demands of modern AI application development.

## Features
- Visual orchestration studio with drag-and-drop interface
- Comprehensive RAG pipeline with multiple indexing and retrieval methods
- Prompt IDE for creating, testing, and optimizing AI prompts
- Enterprise LLMOps with monitoring, logging, and fine-tuning capabilities
- Agentic workflow features with live-editing node debugging
- Support for 10+ commercial models including OpenAI and Anthropic
- Integration with 7+ Model-as-a-Service vendors
- Visual knowledge base management with snippet previews and recall testing
- Support for multimodal capabilities including ASR and rich-text models
- Team collaboration through workspaces and multi-member management

## Integrations
OpenAI, Anthropic, Hugging Face, AWS Bedrock, NVIDIA, Replicate, GroqCloud, together.ai, OpenRouter, Notion, Qdrant, Weaviate, Milvus, Pgvector, Chroma, OpenSearch

## Platforms
WEB

## Pricing
Freemium — Free tier available with paid upgrades

## Links
- Website: https://dify.ai
- Documentation: https://docs.dify.ai
- Repository: https://github.com/langgenius/dify
- EveryDev.ai: https://www.everydev.ai/tools/dify