Dify 1.15.0 (open-source LLM app development platform): api + worker + worker_beat + web + pgvector PostgreSQL (also used as the vector store) + redis + sandbox + ssrf_proxy (squid) + plugin_daemon, fronted by an internal nginx gateway that mirrors the upstream path routing. Closes #88 Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
1.2 KiB
Instructions
First setup
- The first boot runs the database migrations, so it can take 2-4 minutes before the application responds.
- Open your domain: you will be redirected to
/install, where you create the admin (workspace owner) account. - After signing in, go to Settings -> Model Provider and install a model provider plugin (OpenAI, Anthropic, Ollama, etc.) from the marketplace, then configure your API keys.
Notes
-
HTTPS is required for the console login: Dify issues
__Host-/Securesession cookies, so make sure the domain has a certificate (e.g. Let's Encrypt in the domain settings). Over plain HTTP the browser will drop the session cookies and login will not persist. -
The vector store is pgvector, running inside the bundled PostgreSQL instance (no separate Weaviate/Qdrant container needed).
-
All routing (
/console/api,/api,/v1,/files,/e/, ...) is handled by the bundled internal nginx gateway, mirroring the upstream deployment, so the Dify service API is available athttps://<your-domain>/v1. -
Code execution runs in the isolated
sandboxservice, and user-triggered outbound HTTP requests go through thessrf_proxy(squid) service, like in the official deployment.