Dify

Dify is an open source LLM application development platform that takes teams from prototype to production without rebuilding their stack. Where Flowise-langflow-style canvas tools focus on single flows, Dify is a fuller product: it wraps visual workflows, a RAG pipeline, agent capabilities, prompt tooling, and LLM operations into one collaborative workspace with role-based access for teams.

Key features:

• Visual workflow canvas for building and testing AI workflows on a node-based canvas
• Comprehensive model support — hundreds of proprietary and open source LLMs from dozens of providers, including any OpenAI-API-compatible models
• Prompt IDE for crafting prompts, comparing model performance, and adding features like text-to-speech to chat apps
• RAG pipeline covering document ingestion to retrieval, with built-in text extraction from PDFs, PPTs and other formats
• Autonomous agents with a code-execution sandbox that can run commands, install software, and handle files; extendable with tools from the Dify Marketplace, MCP servers, or your own APIs
• LLMOps for monitoring app logs and performance over time, and iterating on prompts, datasets, and models from production data; integrates with Langfuse, Opik and Arize Phoenix
• Backend-as-a-Service — every capability ships with corresponding APIs so Dify can sit behind your own product

Self-hosting options:

• Docker Compose — clone the repo, cd docker && cp .env.example .env && docker compose up -d, then open http://localhost/install (needs 2+ CPU cores and 4+ GiB RAM)
• Source code deployment guide for contributors and custom builds
• Community, Enterprise, and Cloud editions — the managed cloud includes a free sandbox plan; enterprise adds SSO and RBAC

Dify is built with a TypeScript/Next.js frontend and a Python backend. Its Community Edition is free to self-host under the Dify Open Source License (Apache 2.0 with additional conditions restricting multi-tenant hosting and frontend branding removal) — for internal business use it behaves like the free open source alternative to building LLM apps top-to-bottom on proprietary platform APIs.