---
title: "Deploy Plexus"
description: "Unified API gateway for multiple AI providers."
category: "AI/ML"
url: https://railway.com/deploy/plexus
---

# Deploy Plexus

Unified API gateway for multiple AI providers.

**[Deploy Plexus on Railway](https://railway.com/template/plexus)**

- **Creator:** Tom
- **Category:** AI/ML
- **Total deploys:** 4

## Template content

### Plexus https://raw.githubusercontent.com/mcowger/plexus/refs/heads/main/docs/images/plexus_logo_transparent.png

- **Image:** ghcr.io/mcowger/plexus:2026.08.12.1@sha256:91c02e0fe04886216fe0ddc132a6664a49bd269fe73d428b162381ec6dfe1dd3
- **Health check:** /health
- **Public domain:** Yes

## Documentation

# Deploy and Host Plexus on Railway

Deploy Plexus `2026.08.12.1`, a unified gateway for OpenAI, Anthropic, Google Gemini, DeepSeek, Groq, and OpenAI-compatible AI providers.

## About Hosting Plexus

This template runs the official Plexus container as one public service on port 4000. It stores configuration, API keys, usage data, and encrypted provider credentials in SQLite on a persistent Railway volume. `ADMIN_KEY` and `ENCRYPTION_KEY` are generated independently for every deployment.

Open the service domain, enter the generated `ADMIN_KEY`, then configure providers, model aliases, routing, quotas, and scoped API keys in the management dashboard. Keep both generated secrets private and preserve them across redeployments.

## Common Use Cases

- Expose one API surface for multiple AI providers and model protocols.
- Route and fail over requests by availability, latency, performance, or cost.
- Apply per-key rate, token, and spending quotas.
- Track usage and manage provider credentials from a built-in dashboard.
- Proxy OpenAI-compatible chat, response, embedding, transcription, and MCP traffic.

## Dependencies for Plexus Hosting

- Official `ghcr.io/mcowger/plexus:2026.08.12.1` image, pinned by immutable OCI index digest.
- One persistent volume mounted at `/app/data` for the SQLite database.
- Generated 64-character `ADMIN_KEY` and `ENCRYPTION_KEY` values.
- External AI-provider credentials for the providers you choose to enable.

### Deployment Dependencies

- [Plexus 2026.08.12.1 release](https://github.com/mcowger/plexus/releases/tag/2026.08.12.1)
- [Plexus repository](https://github.com/mcowger/plexus)
- [Installation guide](https://github.com/mcowger/plexus/blob/main/docs/INSTALLATION.md)
- [Configuration reference](https://github.com/mcowger/plexus/blob/main/docs/CONFIGURATION.md)
- [API reference](https://github.com/mcowger/plexus/blob/main/docs/openapi/openapi.yaml)

### Implementation Details

The service uses `DATABASE_URL=sqlite:///app/data/plexus.db`, listens on port 4000, and is checked through `/health`. Monthly Railway volume backups are configured. The image is pinned to a stable tag and immutable multi-architecture digest; the template does not use `latest`.

Plexus is ready without provider credentials, but model requests cannot succeed until at least one provider is configured. The persistent database contains sensitive encrypted configuration, so retain the generated encryption key and protect backups.

## Why Deploy Plexus on Railway?

Railway provides managed HTTPS, generated secrets, health checks, persistent storage, backups, and reproducible image deployment for Plexus without requiring a separate database service.


## Similar templates

- [Chat Chat](https://railway.com/deploy/-WWW5r) — Chat Chat, your own unified chat and search to AI platform.
- [stella](https://railway.com/deploy/stella) — Self-host stella with web, API, Postgres, Redis, and object storage.
- [Hermes Agent | OpenClaw Alternative with Dashboard](https://railway.com/deploy/hermes-agent-or-openclaw-alternative-wit) — Self-Hosted Hermes AI Agent for Telegram, Discord & Slack

Open this page in a browser: https://railway.com/deploy/plexus