Deploy and Host Open WebUI on Railway

Open WebUI is a self-hosted, extensible web interface for chatting with large language models. It connects to OpenAI, Anthropic, Ollama, and other OpenAI-compatible providers, and adds RAG document chat, multi-model conversations, web search, and a plugin ecosystem inside one polished, ChatGPT-style UI.

About Hosting Open WebUI

Open WebUI runs as a Python/FastAPI backend serving a compiled Svelte frontend, listening on port 8080. Rather than the single-container SQLite default, this template follows Open WebUI's own scaling guide and deploys four services pre-wired over Railway's private network: Open WebUI itself, a Postgres service (the pgvector/pgvector image, which backs both the application database and the vector store for document embeddings), SearXNG (a privacy-respecting metasearch engine that powers in-chat web search), and Redis (a shared state store). User accounts, chat history, and prompts live in PostgreSQL; uploaded-document embeddings live in pgvector inside that same database (the vector extension is created automatically on first boot); Redis coordinates sessions, sign-out token revocation, and realtime websocket events so the app can scale to multiple replicas; and the Open WebUI service keeps its file uploads on a persistent volume at /app/backend/data. The template wraps all of this with automatic Railway PORT binding, a /health healthcheck, and auto-generated secrets. You add an LLM provider key (or point at Ollama) and you have a private, production-shaped AI chat workspace.

Common Use Cases

Private ChatGPT alternative — a self-hosted chat UI where your conversations and data stay on your own infrastructure
Team AI workspace — shared multi-user access with sign-up controls and role-based permissions
RAG document chat — upload PDFs and documents and ask questions grounded in their content
Ollama frontend — a clean web UI for locally hosted open-source models over Railway's private network
Multi-provider playground — compare GPT, Claude, and open models side by side in one interface

Dependencies for Open WebUI Hosting

PostgreSQL (pgvector) — bundled as the pgvector/pgvector service; stores accounts, chats, and document embeddings
SearXNG — bundled (reuses the protemplate/searxng template) to power in-chat web search
Redis — bundled (redis:7-alpine) for token revocation, sessions, and multi-replica websocket coordination
Persistent volume — Open WebUI file uploads are stored at /app/backend/data
LLM provider — at least one of an OpenAI key, an Anthropic key, or a reachable Ollama instance to actually run inference
Modern browser — the interface is a single-page Svelte app

Deployment Dependencies

Why Deploy Open WebUI on Railway?

Railway is a singular platform to deploy your infrastructure stack. Railway will host your infrastructure so you don't have to deal with configuration, while allowing you to vertically and horizontally scale it.

By deploying Open WebUI on Railway, you get the full stack pre-wired over private networking — the app, a PostgreSQL/pgvector database, a SearXNG search backend, and a Redis state store — with a persistent volume for uploads, automatic PORT binding, a public HTTPS domain, auto-generated secrets, and built-in logs and metrics. Connect it to a self-hosted Ollama service over Railway's private network, or bring your own provider keys, and host your AI chat workspace alongside your databases, agents, and the rest of your stack.