Deploy and Host NextChat on Railway

NextChat (formerly ChatGPT-Next-Web) is an open-source, self-hostable AI chat interface built with Next.js. It supports OpenAI, Azure OpenAI, Google Gemini, Anthropic Claude, and a dozen other LLM providers through a clean, fast web UI — with all conversation data stored locally in the browser by default.

About Hosting NextChat

NextChat is a stateless Next.js application — it proxies API requests to your chosen LLM provider and stores all chat history client-side, so no database is required. Hosting your own instance lets you embed a server-side API key (so users don't need their own), restrict access with a shared password via CODE, control which models are available, and point the app at a custom API endpoint or self-hosted model backend. This Railway template deploys the official Docker image and exposes NextChat on a Railway-provided HTTPS domain immediately after deployment.

Common Use Cases

Team AI access — deploy a shared instance with a server-side API key so teammates can use GPT-4, Claude, or Gemini without each managing their own credentials
Custom model frontend — use NextChat as a web UI for a self-hosted model backend (LocalAI, Ollama, RWKV-Runner) by pointing BASE_URL at your inference server
Controlled access — restrict usage to authorised users via the CODE password mechanism, and lock down model selection with CUSTOM_MODELS and DISABLE_GPT4

Dependencies for NextChat Hosting

An API key for at least one supported LLM provider (OpenAI, Azure, Google Gemini, Anthropic, etc.)

Deployment Dependencies

Implementation Details

Set the following environment variables in your Railway service settings. Only OPENAI_API_KEY (or an equivalent provider key) is required to get started.

Core

OPENAI_API_KEY=sk-xxxxxxxx        # Required if using OpenAI; comma-separate multiple keys
CODE=yourpassword                  # Optional: shared access password for the instance
BASE_URL=https://api.openai.com   # Override to use an OpenAI-compatible proxy or self-hosted endpoint

Provider alternatives (use instead of or alongside OpenAI)

AZURE_URL=https://{resource}.openai.azure.com/openai/deployments/{deployment}
AZURE_API_KEY=azure_key_xxxxx
AZURE_API_VERSION=2024-02-01

GOOGLE_API_KEY=google_key_xxxxx
GOOGLE_URL=https://generativelanguage.googleapis.com/

Access & model control

HIDE_USER_API_KEY=1    # Prevent users from entering their own API key
DISABLE_GPT4=1         # Hide GPT-4 models from the UI
CUSTOM_MODELS=+llama3,+mistral,-gpt-3.5-turbo   # Add (+) or remove (-) models from the list
DEFAULT_MODEL=gpt-4o   # Set the default model on first load

Why Deploy NextChat 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 NextChat on Railway, you are one step closer to supporting a complete full-stack application with minimal burden. Host your servers, databases, AI agents, and more on Railway.