
Deploy New API AI Gateway
New API LLM gateway with PostgreSQL, Redis and a generated root login
Data
Redis
Just deployed
PostgreSQL
Just deployed
/var/lib/postgresql
Gateway
Just deployed
Deploy and Host New API AI Gateway on Railway
New API is an open-source LLM gateway and API management system. It puts OpenAI, Claude, Gemini, Azure, AWS Bedrock, DeepSeek and many other providers behind one OpenAI-compatible endpoint, and adds API keys, per-user quotas, model pricing, usage logs and channel failover, all managed in a web console.
About Hosting New API AI Gateway
The template runs New API v1.0.0-rc.51 from the official image, pinned by digest, with PostgreSQL 18 on a 5 GB volume for every record and Redis 8 for caches and rate limits. New API normally lets the first visitor of a fresh instance create the root administrator. Here the first boot creates root over loopback, before the console is reachable, with a generated password you copy from the New API service variables, and public sign-up starts closed. Refresh cookies are Secure, and New API sees each visitor's real address behind Railway's proxy, so its sign-in limits and token IP allowlists work per visitor.
Why Deploy New API AI Gateway on Railway
Railway gives New API an HTTPS domain for its console and API, private networking to PostgreSQL and Redis, a persistent database volume, generated secrets and health-checked deploys. One service is public; the database and cache are reachable only inside the project.
Common Use Cases
- Give a team one API endpoint and per-person keys for several model providers, with quotas and usage logs
- Fail over between provider accounts or regions and route models to the channels you choose
- Run a metered AI API for customers with your own pricing, groups and top-ups
- Point OpenAI-compatible tools, coding agents and chat apps at a gateway you control
Dependencies for New API AI Gateway Hosting
The deployment uses New API, PostgreSQL and Redis.
Deployment Dependencies
- New API v1.0.0-rc.51 (docs.newapi.ai)
- PostgreSQL 18 with a 5 GB Railway volume, and Redis 8
- API keys for the model providers you add as channels
- Optional: an SMTP provider on port 465 or 587 for email
Template Content

