Dify

Why deploy dify render on Render?

Dify Render is a Render deployment template for self-hosting Dify, an open-source platform for building LLM applications, RAG pipelines, and AI agent workflows. It solves the infrastructure complexity of self-hosting Dify by bundling the API, web frontend, Celery worker, Postgres with pgvector, and Redis into a one-click Blueprint deployment.

This template pre-wires Dify's four services (API, worker, web frontend, Key Value cache) with a managed Postgres instance that has pgvector already enabled for embeddings—no manual Redis provisioning or extension setup required. Celery broker URLs, CORS settings, and cross-service discovery are configured automatically, eliminating the hour-plus of environment variable wiring you'd do yourself. One click deploys the full stack with Render's managed Postgres handling both app data and vector storage, so you skip the typical multi-service orchestration headache.

Architecture

What you can build

After deploying, you'll have a self-hosted Dify instance where you can build and run LLM-powered chatbots, RAG pipelines, and multi-step agent workflows through a visual console. The setup includes managed Postgres with pgvector for storing embeddings alongside app data, plus a Celery worker for handling async jobs like document indexing. You'll need to add your own LLM provider keys (OpenAI, Anthropic, etc.) after the initial deploy to start building apps.

Key features

• Managed pgvector storage: Embeddings and app data share a single Postgres 16 instance with the vector extension auto-enabled on first deploy.
• Pre-wired Celery worker: A dedicated background worker connects to Render Key Value as the Celery broker for async jobs without manual Redis setup.
• Official Docker images: Deploys pinned langgenius/dify-api and langgenius/dify-web images directly, avoiding multi-GB builds on each deploy.
• Auto-configured service URLs: CORS, console, and API URLs are injected automatically from Render service discovery so cross-origin requests work immediately.
• Persistent file storage: A 10 GB disk mount stores uploads by default, with optional S3 configuration for durable multi-instance storage.

Use cases

• Product team prototypes an internal customer support chatbot without managing infrastructure
• Developer builds RAG search over private company documentation using pgvector
• Startup chains multiple LLM providers into automated workflow pipelines
• Engineer deploys self-hosted AI gateway to compare OpenAI and Anthropic responses

What's included

Next steps

1. Open the dify-web URL at /install and create your admin account — You should see the Dify setup wizard and be able to set your email and password to access the main console
2. Configure a model provider in Settings → Model Provider by adding an OpenAI or Anthropic API key — The provider should show a green connected status and appear in the model dropdown when creating apps
3. Test the full stack by creating a simple chatbot app and sending a message — You should receive an LLM response within a few seconds, confirming the API, worker, database, and Key Value are all communicating correctly

Resources

• Dify Github