# Dify

> Dify is a platform for building LLM apps: AI workflows on a visual canvas, agents with tools and MCP servers, a RAG knowledge base for PDFs and other documents, and an API for every app. Workflows can run on a schedule or from a webhook. It supports hundreds of models, including local ones through the Ollama plugin. Self-host with Docker Compose (2 cores, 4 GB RAM minimum) or use Dify Cloud. The licence is Apache 2.0 plus conditions: no multi-tenant service without written permission, and the logo and copyright in the web frontend must stay.

- Repository: https://github.com/langgenius/dify
- Official docs: https://docs.dify.ai
- Checked: 2026-10-21

## What it replaces
- Zapier: Scheduled automations, AI agent workflows, Webhooks (partial). Strong for AI workflows with schedule and webhook triggers; fewer ready-made app integrations, added through marketplace plugins.
  - Use it instead of paying for Zapier? Personal use: Yes. Business: Partly (Commercial use with conditions: Commercial use allowed, but a multi-tenant service needs written permission and the web f…).
- Make: Scheduled automations, AI agent workflows, Webhooks (partial). Visual canvas built around LLM steps rather than general app-to-app automation.
  - Use it instead of paying for Make? Personal use: Yes. Business: Partly (Commercial use with conditions: Commercial use allowed, but a multi-tenant service needs written permission and the web f…).
- ChatGPT Plus: Chat with your documents (partial). Upload documents to a knowledge base and chat with them through an app you set up, using your own or local models.
  - Use it instead of paying for ChatGPT Plus? Personal use: Yes. Business: Partly (Commercial use with conditions: Commercial use allowed, but a multi-tenant service needs written permission and the web f…).

## Facts, with sources

| Question | Status | Answer | Source | Checked |
|---|---|---|---|---|
| Runs on my machine? | Confirmed | Self-hosted with Docker Compose on your machine or server; Dify Cloud also offered | https://github.com/langgenius/dify#using-dify | 2026-10-21 |
| Code licence | Partial | Dify Open Source License: Apache 2.0 with extra conditions (no multi-tenant service without permission; keep the logo and copyright in the frontend); not an OSI licence | https://github.com/langgenius/dify/blob/main/LICENSE | 2026-10-21 |
| Model weights licence | Confirmed | No separate model weights | https://github.com/langgenius/dify#readme | 2026-10-21 |
| Commercial use | Partial | Commercial use allowed, but a multi-tenant service needs written permission and the web frontend's logo and copyright may not be removed | https://github.com/langgenius/dify/blob/main/LICENSE | 2026-10-21 |
| Paid services needed? | Partial | Not required with local models via the Ollama plugin; cloud model APIs, Dify Cloud and Dify Enterprise are optional | https://marketplace.dify.ai/plugin/langgenius/ollama | 2026-10-21 |
| Hardware | Confirmed | Minimum 2 CPU cores and 4 GiB RAM; on macOS the docs ask for a Docker VM with 2 vCPUs and 8 GiB | https://docs.dify.ai/en/self-host/deploy/quick-start/docker-compose | 2026-10-21 |
| Operating systems | Confirmed | Docker on Linux, macOS 10.14+ and Windows (with WSL 2) | https://docs.dify.ai/en/self-host/deploy/quick-start/docker-compose | 2026-10-21 |
| Setup effort | Confirmed | Medium; clone the repo, copy .env and run docker compose up | https://github.com/langgenius/dify#quick-start | 2026-10-21 |

GitHub: 159161 stars, 25155 forks, last push 2026-10-22, archived: no.
---
Source: RepoQuarry (https://repoquarry.dev/projects/dify). Data licence: CC BY 4.0. Quotes are from project documentation and belong to their authors.