Deploy GoModel

Deploy and Host GoModel with Railway

GoModel is a lightweight, open-source AI gateway with OpenAI-compatible and Anthropic-compatible APIs. Connect applications to multiple model providers through one endpoint, with streaming, model routing, and a built-in dashboard. Manage provider access, create application API keys, and track token usage and estimated costs in one place.

About Hosting GoModel

Choose this template if… you want a separate database, shared response caching, and the option to run multiple GoModel instances. For a simpler single-service setup, choose GoModel Lite.

The template includes GoModel's API and dashboard, PostgreSQL for persistent data, and Redis for caching, already connected over Railway's private network. HTTPS and credentials are configured automatically. No provider key is needed to deploy; bring your own provider access to make model requests. This template starts with one instance per service, not a highly available cluster. Configure database backups before production use.

Common Use Cases

• Connect AI applications to multiple providers through a consistent API.
• Route agents and coding tools through your own gateway.
• Track token usage and estimated costs by application.
• Reuse identical model responses through shared caching.

Dependencies for GoModel Hosting

• PostgreSQL and Redis with persistent volumes; included and configured.
• Access to a supported provider or a reachable self-hosted model endpoint. Model hosting and inference credits are not included; provider charges are separate from Railway hosting.

Deployment Dependencies

• GoModel source and official Docker image
• Production and scaling guidance
• Response caching and Railway backups

Implementation Details

Get started after deployment

1. Find GoModel's public domain under Settings → Networking in Railway and open https://YOUR_DOMAIN/admin/dashboard.
2. Sign in with GOMODEL_MASTER_KEY from GoModel's Variables tab in Railway.
3. Open Providers → Add Provider, choose your provider type, paste its API key, and click Save.
4. Open Playground. Once model discovery finishes, select a chat model, type “Say hello”, and click Send to make your first request.

Dashboard-added provider keys are stored as plaintext; prefer Railway variables for production. Audit-body logging is disabled, but Redis caches generated responses with a one-hour TTL.

For API integration, routing, and other advanced usage, see the GoModel documentation.

Why Deploy GoModel on Railway?

Run your gateway, database, and cache in one project with HTTPS, private networking, logs, and resource metrics. Railway handles hosting so you can focus on connecting your applications.