Dify v1 AI App Studio

Dify is an open-source platform for building LLM apps: chatbots, RAG knowledge bases, agents and workflows. This template deploys Dify 1.17 as a fifteen-service stack behind a single HTTPS gateway, with generated secrets.

What you get

  • Gateway, web console, API, workers, plugin daemon and code sandbox as separate services
  • Postgres, Redis and Weaviate (vector store), with volumes on the stateful services
  • Railway Bucket for uploaded files
  • Two Squid proxies that block requests to private networks
  • All keys and passwords generated at deploy; no input required

Good for

  • Internal knowledge assistants that answer from your own documents (RAG)
  • Embeddable chatbots backed by an LLM provider you choose
  • Multi-step AI workflows exposed as an API
  • Tool-using agents with sandboxed code execution

Services in this template

  • API · image langgenius/dify-api:1.17.1
  • Worker · image langgenius/dify-api:1.17.1
  • SSRF Proxy · from baranberkay96/dify-railway
  • Web · image langgenius/dify-web:1.17.1
  • Agent SSRF Proxy · from baranberkay96/dify-railway
  • Postgres · image ghcr.io/railwayapp-templates/postgres-ssl:18
  • API WebSocket · image langgenius/dify-api:1.17.1
  • Redis · image redis:8.2
  • Weaviate · image semitechnologies/weaviate:1.39.11
  • Sandbox · image langgenius/dify-sandbox:0.2.15
  • Worker Beat · image langgenius/dify-api:1.17.1
  • Agent Sandbox · from baranberkay96/dify-railway
  • Plugin Daemon · image langgenius/dify-plugin-daemon:0.6.10-local
  • Agent Backend · image langgenius/dify-agent-backend:1.17.1
  • Gateway · from baranberkay96/dify-railway

Questions people ask

How do I log in the first time?

Wait for the API logs to show the migrations finished (1 to 3 minutes). Open the Gateway URL at /install, enter INIT_PASSWORD from the API service variables, then create the owner account.

Is data kept across redeploys?

Yes. Postgres, Redis, Weaviate, the plugin daemon, the code sandbox and the agent sandbox use volumes, and uploaded files go to a Railway Bucket.

How do I upgrade Dify?

Images are pinned. Read the Dify release notes, change all 1.17.1 tags together, and update the plugin daemon, sandbox and Weaviate tags to match that release. Back up Postgres and the Weaviate volume before major upgrades.

What drives the cost?

Fifteen services run in one project, so memory across the services is the main driver. The worker and plugin daemon grow under load. The template notes recommend at least the Hobby plan.

Are there licence restrictions?

Dify uses the Dify Open Source License (Apache 2.0 with additional conditions). Running it for your own organisation is permitted. Operating it as a multi-tenant service for third parties, or removing the console logo and copyright, requires a commercial licence from LangGenius.

Full template README

Deploy and Host Dify with Railway

Dify is an open-source platform for building LLM applications: chatbots, RAG knowledge bases, agents and multi-step workflows, designed in a visual editor and published as web apps or APIs. This community template runs the complete Dify 1.17 stack on Railway behind a single HTTPS URL.

About Hosting Dify

Self-hosting Dify means running far more than one container. The console (Next.js) talks to a Flask API; Celery workers index documents and execute workflows; a beat scheduler fires timed jobs; a plugin daemon installs and runs model/tool plugins; a code sandbox executes user code; and the new Agent backend drives sandboxed agent shells. Postgres stores metadata, Redis carries queues and locks, Weaviate holds embeddings, and uploaded files need object storage. This template wires all of it with generated secrets, private networking, a Railway Bucket for files, persistent volumes for stateful services, and an nginx gateway that reproduces Dify's own routing on one domain.

Common Use Cases

  • Internal knowledge assistants that answer from your PDFs, wikis and policies (RAG with citations)
  • Customer-facing chatbots and embeddable chat widgets backed by any LLM provider
  • Multi-step AI workflows (classification, extraction, summarisation) exposed as an API for your products
  • Tool-using agents that browse, call APIs and run code in a sandbox
  • A shared, governed LLM workspace where teams build and publish AI apps without writing backend code

Dependencies for Dify Hosting

  • Dify API, Worker, Worker Beat and WebSocket (langgenius/dify-api), Web (langgenius/dify-web)
  • Dify Plugin Daemon, Code Sandbox, Agent Backend and Agent Sandbox
  • PostgreSQL (metadata), Redis (queues, cache, locks), Weaviate (vector store)
  • Railway Bucket (S3-compatible file storage)
  • nginx gateway and two Squid SSRF proxies (built from this repository)

Deployment Dependencies

Implementation Details

Service Image Role Public
Gateway nginx 1.30 (repo build) Single origin; routes /console/api, /api, /v1, /files, /mcp, /triggers → API, /socket.io/ → WebSocket, /e/ → Plugin Daemon, / → Web yes
Web dify-web:1.17.1 Console and published web apps no
API dify-api:1.17.1 REST API, runs DB migrations no
API WebSocket dify-api:1.17.1 Real-time collaborative editing no
Worker dify-api:1.17.1 Celery: indexing, workflows, mail, plugins no
Worker Beat dify-api:1.17.1 Celery beat: scheduled tasks and triggers no
Plugin Daemon dify-plugin-daemon:0.6.10-local Installs and runs plugins (volume) no
Sandbox dify-sandbox:0.2.15 Executes Code nodes (volume for extra packages) no
SSRF Proxy Squid (repo build) Egress filter blocking private networks no
Agent Backend dify-agent-backend:1.17.1 Agent run orchestration no
Agent Sandbox dify-agent-local-sandbox:1.17.1 (repo build) Agent shell workspaces and home snapshots (volume) no
Agent SSRF Proxy Squid (repo build) Egress filter for the agent sandbox no
Weaviate weaviate:1.39.11 Vector store (volume) no
Postgres / Redis Railway base images Metadata / queues and cache no
Bucket Railway Bucket Uploaded files and generated assets —

First login

  1. Wait until API logs show the migrations finished (1–3 minutes on first boot).
  2. Open the Gateway URL and go to /install. Dify asks for the init password: copy INIT_PASSWORD from the API service variables.
  3. Create the owner account, then add a model provider under Settings → Model Provider (installs a plugin from the Dify marketplace).

Custom domain: add it to the Gateway service, then set DIFY_PUBLIC_URL on API to https://your.domain and NEXT_PUBLIC_SOCKET_URL on Web to wss://your.domain, and redeploy.

Email (optional): set MAIL_TYPE (smtp or resend), MAIL_DEFAULT_SEND_FROM and the SMTP or Resend variables on API, Worker and Worker Beat to enable invitations and password resets.

Scaling workers: increase replicas on Worker (stateless) or raise CELERY_WORKER_AMOUNT. Keep Worker Beat at one replica. For more API throughput raise SERVER_WORKER_AMOUNT on API. Keep total database connections below Postgres' limit of 100.

Pinning and upgrading: every image is pinned. To upgrade, read the Dify release notes, then change all 1.17.1 tags together (API, API WebSocket, Worker, Worker Beat, Web, Agent Backend and the Agent Sandbox Dockerfile) and update Plugin Daemon, Sandbox and Weaviate to the tags in that release's docker/docker-compose.yaml. The API service applies database migrations on start. Back up Postgres (and the Weaviate volume) before major upgrades.

License note: Dify is distributed under the Dify Open Source License (Apache 2.0 with additional conditions). Running it for your own organisation is permitted; operating it as a multi-tenant service for third parties, or removing Dify's logo/copyright from the console, requires a commercial license from LangGenius. This is a community template and is not affiliated with LangGenius.

Why Deploy Dify on Railway?

Railway runs all fifteen Dify services in one project with private networking, managed TLS, persistent volumes and an S3-compatible bucket, so you get a production-shaped Dify without maintaining servers or Docker Compose. Scale workers with a slider, roll back a deploy in one click, and pay for the resources the stack actually uses.

Guides for this template

View Dify v1 AI App Studio on railway.com

Similar templates

TemplateCategoryDeploymentsDeploys
9Router
Self-hosted AI router for Claude, Gemini, Cursor & OpenAI tools.
AI/ML9,734
flowise-railway
Flowise - low-code LLM apps builder
AI/ML8,512
LibreChat
Open-source ChatGPT clone: multi-model, agents, search, multi-user auth
AI/ML6,762
Hermes Agent by Nous Research
Template for Hermes Agent by Nous Research
AI/ML6,413
OpenClaw
[Homebrew support] The AI that actually does things
AI/ML5,167
Hermes Agent
Self Host Hermes Agent. Autonomous AI agent w/ self-improving, tools & more
AI/ML4,092