Dify

A platform for building AI apps, agents and retrieval pipelines on a visual canvas, which you can run in Dify's cloud or on your own servers.

160k stars, 25.3k forks Python On GitHub since April 2023

Visit website View on GitHub How to run it

What is Dify?

Dify lets teams assemble LLM applications from building blocks: workflows on a visual canvas, retrieval over your documents (RAG), agents with tools, and management of the models and providers behind them. Finished apps are published as chat interfaces or APIs. The self-hosted version starts with Docker Compose and needs at least 2 CPU cores and 4 GB of RAM.

Good fit if

  • You want product and ops people to build AI assistants and document Q&A without writing much code.
  • You need to switch between model providers, including local ones, without rebuilding the app.
  • You prefer to keep prompts, documents and logs on your own infrastructure.

Look elsewhere if

  • You plan to sell a hosted, multi-tenant service built on it: the license forbids that without written permission. Langflow (MIT) has no such limit.
  • You only need classic app-to-app automation without AI: n8n or Activepieces fit better.

Before you adopt it

Dify is not under plain Apache 2.0 but under a modified version with extra conditions, so it is not an OSI-approved open-source license. Running a multi-tenant environment needs a commercial license, and you may not remove the Dify logo or copyright notices from its web interface. Enterprise code lives in a separate directory of the repository. The default stack runs about 15 containers.

Drafted with AI from the project's README and repository data, then checked against that data on 30 October 2026. Report a correction

Dify health: is it still maintained?

84 of 100, actively developed

  • Activity
    2,489 commits in the last 90 days, the latest on 1 November 2026.
    30 / 30
  • Contributors
    36 people wrote the latest 100 commits; the most active wrote 20% of them.
    20 / 20
  • Releases
    13 releases in the last 12 months; the latest, 1.18.0, on 23 October 2026.
    19 / 20
  • Replies to issues
    Too few recent issues from outside the team to measure, so this part gets a neutral score.
    10 / 20
  • Security practices
    No OpenSSF Scorecard for this repository yet, so this part gets a neutral score.
    5 / 10

Read from GitHub on 1 November 2026. How the score works

Dify license: is it open source?

Dify uses Apache-2.0 with additional conditions. It is source-available, not an OSI-approved open-source license: the code is public, but the license restricts how you may use it, usually commercially. Some features sit under a separate paid license inside the same repository.

  1. April 2023First published on GitHub.
  2. TodayApache-2.0 with additional conditions, with paid parts: directory api/enterprise/.

How to run Dify yourself

Docker
Dockerfile or compose file in the repository
Compose file
docker/docker-compose.yaml: 15 services (postgres, mysql, redis)
Setup effort
4 of 5, judged from the compose file
Kubernetes
No Helm chart in the repository
Desktop and mobile
Builds for Linux, macOS, Windows, including ARM in GitHub releases
Interface languages
22 languages

Install instructions: dify.ai or the README on GitHub.

Alternatives to Dify

  • n8nFair-code workflow automation platform with native AI capabilities. Combine visual building with custom… 84 health · Sustainable Use License · 209k stars
  • Trigger.devTrigger.dev – build and deploy durable AI agents and workflows 84 health · Apache-2.0 · 16.7k stars
  • Apache AirflowApache Airflow - A platform to programmatically author, schedule, and monitor workflows 87 health · Apache-2.0 · 47.3k stars
  • KestraOrchestrate everything, everywhere. The open-source control plane for workflows across data, infrastructure,… 85 health · Apache-2.0 · 30.1k stars
  • PrefectPrefect is a workflow orchestration framework for building resilient data pipelines in Python. 80 health · Apache-2.0 · 24.1k stars

All 33 projects in Workflow automation

Questions about Dify

Is Dify open source?

Not by the OSI definition. Dify uses Apache-2.0 with additional conditions, which makes the code public but restricts how it may be used.

Is Dify still maintained?

Its health score is 84 of 100 (actively developed). It had 2,489 commits in the last 90 days, the latest on 1 November 2026, and its latest release came out on 23 October 2026.

How do I run Dify?

You can run it with Docker, as an app for Linux, macOS, Windows, including ARM. The README on GitHub has the exact steps.

What are the alternatives to Dify?

Close alternatives in the Workflow automation category include n8n, Trigger.dev, Apache Airflow.