Deploy and Host Sirene Inference on Railway

Sirene Inference is the worker that runs text-to-speech generation for Sirene, an open-source TTS studio with voice cloning and voice design. This template deploys one worker that registers itself with your Sirene server on first start, so it shows up in your inference server list without any manual setup.

About Hosting Sirene Inference

The worker is a FastAPI service packaged as a Docker image. It listens on the port Railway provides, exposes a /health endpoint that Sirene uses to monitor it, and keeps downloaded models and lazily installed backends on the attached volume so they survive redeploys. At startup it calls your Sirene server with its public URL and its own auth token; Sirene upserts the entry by URL, so redeploys never create duplicates. Railway offers CPU only, which suits light models such as Kokoro or Piper; larger models need a GPU host. Your Sirene server must be reachable from the internet.

Common Use Cases

Offload speech generation from a small Sirene server to a separate worker
Try Sirene's multi-server failover without a second machine
Run light TTS models close to your users

Dependencies for Sirene Inference Hosting

A running Sirene server with a public URL
A registration token from Sirene, under Administration → Inference servers → Add server (valid one hour)

Implementation Details

Set SIRENE_URL and SIRENE_REGISTRATION_TOKEN when deploying; every other variable is generated or derived from Railway. Source and documentation: https://github.com/KevinBonnoron/sirene

Why Deploy Sirene Inference 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 Sirene Inference 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.