One-click deployment of OmniRoute, the open-source AI gateway, with the Deploy to NEXUS AI button

How we deploy OmniRoute, an MIT-licensed gateway for hundreds of AI providers, with the Deploy to NEXUS AI button. No config, no servers, Free plan included

If you use Claude Code, Codex, Cursor, or OpenCode every day, you have probably felt this: one provider hits a rate limit in the middle of a task, you switch keys by hand, and a different tool needs a different config for every provider.

An AI gateway fixes that. You point every tool at one OpenAI-compatible endpoint, and the gateway decides which provider serves each request.

OmniRoute is one of the most capable open-source gateways available. It is MIT-licensed, maintained by Diego Souza, and it gives you:

The catch with any self-hosted gateway is the hosting. You need a server, HTTPS, a place for secrets, and time to set it all up. We wanted to remove that step.

One click, a live gateway

We tested OmniRoute end to end on NEXUS AI and it deploys with zero configuration. It generates its own secrets on first start, so there is nothing to fill in.

Deploy it here: nexusai.run/deploy?repo=https://github.com/diegosouzapw/OmniRoute

Here is what happens when you click:

  1. NEXUS AI reads the OmniRoute repository and shows you exactly what will be created.
  2. You sign in (or create a free account) and land back on the same page.
  3. You pick a name and click Deploy.
  4. NEXUS AI builds OmniRoute from its own Dockerfile and gives it a public HTTPS URL.

The first build takes several minutes. After that, you have your own gateway at https://<your-app>.nexusai.run.

Finish setup in four steps

  1. Open your URL. The setup wizard asks for a one-time setup token.
  2. Copy the token from the Logs tab of your deployment in NEXUS AI, paste it, and choose your password.
  3. Connect a provider under Providers, then copy an API key from Endpoints.
  4. Point your tools at https://<your-app>.nexusai.run/v1.

That is it. Claude Code, Codex, Cursor, and any OpenAI-compatible client now go through your own gateway, with fallback across every provider you connected.

Secure from the first minute

A public gateway with no lock is a real risk. OmniRoute handles this well, and it is one of the reasons we picked it:

What it costs

OmniRoute runs on the NEXUS AI Free plan. Connect free-tier providers and your gateway costs nothing to run. If you outgrow Free, the same deployment moves to a paid plan with no changes.

One thing to know: OmniRoute keeps its database and generated secrets inside the container. Stopping and starting the deployment keeps them. A full rebuild starts fresh, so connect your providers again after a rebuild.

Why this matters for open source

The Deploy to NEXUS AI button works for any public GitHub repository. Maintainers add one line to their README, and anyone can launch their own running copy, with databases, workers, and required settings described in an optional nexus.json file.

For a project like OmniRoute, that removes the gap between "this looks useful" and "it's running and my tools are using it." That gap is where most people give up on self-hosting.

If you maintain an open-source app, you can add the same button to your repo in five minutes: nexusai.run/kb/deploy-button

Try it

About NEXUS AI

NEXUS AI is an agentic AI app builder and full-stack deployment platform. Explore the AI App Builder, learn more on the About page, read the documentation, or contact the team through nexusai.run/contact.

Start for free · Read the documentation · About NEXUS AI · Contact