Dify on your own server: AI apps, agents and workflows
Build chatbots, agents, knowledge bases and AI workflows in a visual studio, connect any model provider or your own models, and publish every app as an API. Run the official Docker Compose stack on a HyperDC server that meets Dify’s requirements, preinstalled as an app option or set up with our guide.
- Visual workflows, agents and RAG knowledge bases
- Hosted model APIs or your own models through Ollama
- Every app also published as an API
- Sized from Dify’s minimum of 2 cores and 4 GiB
- StackThe stack runs the API, workers, the web app, a plugin daemon and a sandbox, with PostgreSQL, Redis and Weaviate.
- Python and Next.js, Docker Compose
- Default ports
- 80 and 443 (nginx)
- Minimum
- 2 CPU cores, 4 GiB RAM
- Your data
- PostgreSQL, Redis, vector store, files
- Official docs
- docs.dify.ai
Build with
- Workflows
- Chatflow
- Agents
- Knowledge bases
- MCP
- Plugins
- Ollama
- OpenAI-compatible APIs
Facts from the project’s official website, documentation and repository, checked in October 2026.
Plans are being prepared
We are preparing ready-to-use plans for Dify. Tell us how you will use it and how many users you expect, and we will reply with a server that fits. You can also start today on a Linux VPS and install it with our guide.
Which server size fits?
Starting points for vCPU, memory and disk. Grow the server when your data and users grow.
| Feature |
Try it
One builder, hosted model APIs
|
Recommended Team
Several apps and knowledge bases
|
Production
Many users, large document sets
|
|---|---|---|---|
| vCPUVirtual processor cores of the server. | 2 | 4 | 8 |
| MemoryMemory for the app, its database and the operating system. | 4 GB | 8 GB | 16 GB |
| DiskDocker images, PostgreSQL, vector store and uploaded files. | 40 GB | 80 GB | 160 GB+ |
| ModelsWhere the language model runs. | Hosted APIs | APIs or Ollama on a second server | Your own models on a GPU server |
| Server type | Linux VPS | VPS or VDS | VDS or dedicated |
-
Try it
One builder, hosted model APIs
- vCPUVirtual processor cores of the server.
- 2
- MemoryMemory for the app, its database and the operating system.
- 4 GB
- DiskDocker images, PostgreSQL, vector store and uploaded files.
- 40 GB
- ModelsWhere the language model runs.
- Hosted APIs
- Server type
- Linux VPS
-
Recommended
Team
Several apps and knowledge bases
- vCPUVirtual processor cores of the server.
- 4
- MemoryMemory for the app, its database and the operating system.
- 8 GB
- DiskDocker images, PostgreSQL, vector store and uploaded files.
- 80 GB
- ModelsWhere the language model runs.
- APIs or Ollama on a second server
- Server type
- VPS or VDS
-
Production
Many users, large document sets
- vCPUVirtual processor cores of the server.
- 8
- MemoryMemory for the app, its database and the operating system.
- 16 GB
- DiskDocker images, PostgreSQL, vector store and uploaded files.
- 160 GB+
- ModelsWhere the language model runs.
- Your own models on a GPU server
- Server type
- VDS or dedicated
Dify’s documentation asks for at least 2 CPU cores and 4 GiB of RAM for the Docker Compose stack. Larger columns are conservative starting points for more apps, users and documents; local models need their own server.
What teams build with Dify
From a support bot over your documents to multi-step automations with tools.
Knowledge bases and RAG
Upload PDFs, slides and web pages, let Dify split and index them, and answer questions with sources.
Workflows and agents
Chain model calls, code, tools and conditions on a canvas, or give an agent tools and MCP servers to work with.
An API for every app
Every app you build is also a REST API, so your website or product can use it; apps can also be published as MCP servers.
Your data stays yours
Prompts, files and databases stay on a server you control, in the location you choose, instead of a shared SaaS account.
Full root access
Install what the app needs, change any setting and run more services next to it. Nothing is locked behind a panel.
App option or step-by-step guide
Order the server with the app installed as an option, or set it up yourself on a clean Linux server with our guide.
Near your users
Choose a data center in the United States, Europe or Asia. The order form estimates the latency from where you are to each location.
Backups on VPS and VDS
Free weekly and monthly backups come with VPS and VDS plans. Keep your own copies of databases and app data as well.
From order to first login
Order the app preinstalled on your server, or install it yourself with our guide.
-
Pick the server
Choose a size from the table above and the data center closest to the people who will use the app.
-
Add Dify
Select Dify as an app option when you order, or install it on a clean Ubuntu or Debian server with our guide.
-
Point a domain and enable HTTPS
Create a DNS record such as app.example.com for the server and put a reverse proxy with a free Let’s Encrypt certificate in front of the app.
-
Create the admin account
Open /install, create the first admin, then add your model provider keys or your Ollama endpoint under Settings.
Step-by-step setup guides
Install, secure and update the app with our guides, written for current Ubuntu and Debian releases.
-
How to self-host Dify with Docker Compose and HTTPS
Clone a Dify release, replace the default passwords and keys in .env, bind the bundled Nginx to localhost behind Caddy, protect the first-run page and learn the backup and upgrade routine.
45 min Intermediate
Related solutions
Frequently asked questions
What do I need to run Dify?
A Linux server with Docker and Docker Compose 2.24 or later, and according to Dify’s documentation at least 2 CPU cores and 4 GiB of RAM. The official compose file starts every service, including PostgreSQL, Redis and the Weaviate vector store.
Can Dify use my own models?
Yes. Besides hosted providers, Dify connects to self-hosted models, for example an Ollama server or any OpenAI-compatible endpoint such as vLLM or LocalAI. Run large models on a separate GPU server so Dify itself stays responsive.
Which passwords must I change after installing?
The example .env file ships with fixed development values for the database, Redis, the sandbox, the plugin daemon and the vector store. Replace every one of them with your own long random values before you go live, and set your own SECRET_KEY.
Is there a default admin login?
No. After the stack starts, you create the first admin account on the /install page. Do this right away, before you share the address.
How do I update Dify?
Back up the docker/volumes folder and your .env file, pull the new release, compare your .env with the new .env.example, and run docker compose up -d. Dify releases new versions often, so read the release notes first.
Where is my data stored?
In the volumes of the compose stack on your server: PostgreSQL for apps and conversations, the vector store for knowledge-base embeddings and the storage folder for uploaded files. Nothing leaves the server except the calls you make to model providers.
Build your AI apps on your own server
Tell us the app, how many people will use it and where they are, and we will suggest a server for it.