@ffxnexus/nexus

Run the Nexus LLM gateway on your machine. One command, no database to install, no config file to write.

npx -y @ffxnexus/nexus

Gateway on http://localhost:8080, console on http://localhost:8081. Open the console, create an account, add a provider key, mint a virtual key, then point any OpenAI client at it:

export OPENAI_BASE_URL=http://localhost:8080/v1
export OPENAI_API_KEY=nxs_live_...

What this package does

It downloads the nexus binary for your platform from the matching GitHub release, verifies it against the published checksums.txt, caches it under ~/.nexus/bin/<version>/, and runs it as nexus serve --local.

Local mode starts a private Postgres next to the gateway so the console can actually store the provider keys and virtual keys you create. Everything — the database, the encryption key — lives in ~/.nexus and survives restarts.

Any arguments you pass go straight to the binary, so npx @ffxnexus/nexus serve runs without the local database and npx @ffxnexus/nexus migrate reaches the migration command.

Requirements

macOS or Linux, on x86-64 or arm64, with Node 18+ and tar. On any other platform, use the container image:

docker run -p 8080:8080 -p 8081:8081 -v "$PWD/data:/app/data" ghcr.io/fun-fx/ffx_nexus

Environment

| Variable | Effect | | --- | --- | | NEXUS_LOCAL_STATE_DIR | Where the binary cache, database and master key live. Default ~/.nexus. | | NEXUS_RELEASE_BASE_URL | Fetch the binary from a mirror instead of github.com. | | NEXUS_GATEWAY_ADDR / NEXUS_CONSOLE_ADDR | Listen addresses. Default :8080 / :8081. |

Local mode is for one machine. For a cluster, use the Helm chart — see the deployment docs.