Use it

One server, one endpoint: POST /v1/systemone. Everything below assumes the default port 21160.

Run it

macOS / Linux
npx git+https://github.com/muthuishere/openjevx.git
openjevx

or download a package for your platform from the latest release:

macOS
curl -L -O https://github.com/muthuishere/openjevx/releases/download/v0.5.11/openjevx-darwin-arm64.tar
curl -L -O https://github.com/muthuishere/openjevx/releases/download/v0.5.11/openjevx-model-0.5.2.tar.gz
tar -xf openjevx-darwin-arm64.tar && tar -xzf openjevx-model-0.5.2.tar.gz && ./openjevx
Linux
curl -L -O https://github.com/muthuishere/openjevx/releases/download/v0.5.11/openjevx-linux-amd64.tar
curl -L -O https://github.com/muthuishere/openjevx/releases/download/v0.5.11/openjevx-model-0.5.2.tar.gz
tar -xf openjevx-linux-amd64.tar && tar -xzf openjevx-model-0.5.2.tar.gz && ./openjevx

Linux on ARM (arm64): use openjevx-linux-arm64.tar instead.

Windows: download and run openjevx-windows-amd64.zip with openjevx-model-0.5.2.tar.gz unpacked into the same folder.

Docker

Docker
docker run -d -p 127.0.0.1:21160:21160 \
  -e OPENJEVX_PASSWORD=<12+ characters> -e OPENJEVX_API_KEY=<16+ characters> \
  ghcr.io/deemwar-products/openjevx:v0.5.9

The prebuilt image is for linux/amd64 and linux/arm64. To build it yourself:

Docker (from source)
git clone https://github.com/muthuishere/openjevx.git && cd openjevx
OPENJEVX_PASSWORD=<12+ characters> OPENJEVX_API_KEY=<16+ characters> docker compose up -d --build

The image has no default credentials and refuses to start without both. It runs as a non-root user.

Wire it into jevx

shell
jevx profile add openjevx http://127.0.0.1:21160/v1/systemone --model openjevx
jevx profile use openjevx

jevx is "Is this urgent?" < email.txt
jevx pick "Which team?" web=frontend api=backend billing=money --in "charged twice"
jevx ask --noul out_of_stock="Is the item out of stock?" --in '{"stock_units": 0}'

A server that listens beyond 127.0.0.1 needs an API key: add --header 'Authorization: Bearer $OPENJEVX_API_KEY' to jevx profile add.

Config

openjevx.json next to the binary:

openjevx.json
{
  "listen": "127.0.0.1:21160",
  "device": "auto"
}

device is auto (the first GPU provider that loads: CUDA, CoreML on macOS, DirectML on Windows; else CPU), cpu, or gpu (no fallback).

Credentials

Dashboard and stats

Open http://127.0.0.1:21160/ while it runs — request counts, input tokens, latency p50/p95/p99, errors, recent requests.

Downloads

Hugging Face

huggingface.co/muthuishere/openjevx — full precision + quantized weights