ONNX Runtime

Adds a local ONNX (onnxruntime) model runtime to IVYX Studio — install it if you want to run ONNX-format models on your own machine for chat, embeddings, and reranking. It has no panel of its own; once installed, ONNX models become runnable.

What you can do

Once a model is running, this runtime answers chat requests plus embeddings, reranking, speech-to-text, and vision — through the model's OpenAI-compatible server.

Getting started

1. Point llm.onnx.serverPath at a script or binary that starts an OpenAI-compatible onnxruntime(-genai) server (exposing /v1 chat, embeddings, and rerank routes).
2. Start an ONNX model from the Models panel. The runtime launches the server and serves the model.

Starting and stopping it from an agent

The server lifecycle is also available as three capabilities, so an assistant or a flow can bring a local model up without going through the Models panel.

Capability	What it does
llms.onnx.server.start	Starts the configured server and points local model calls at it. Starting a process is high risk, so this one asks for approval every time.
llms.onnx.server.stop	Stops the server and releases the local address, so later calls report that nothing is running instead of failing against a dead one.
llms.onnx.server.status	Reports whether a server is up, and where. Reads only.

Starting twice returns the server that is already running rather than putting a second one on the same port.

Requirements

• Desktop (Electron) app — the local server can't run in the browser version.
• An OpenAI-compatible onnxruntime server binary or script.

Settings

Setting	Default	Description
llm.onnx.serverPath	—	Command that starts the OpenAI-compatible onnxruntime(-genai) server.
llm.onnx.serverArgs	—	Arguments for that command; supports {model}, {port}, and {host} placeholders.
llm.onnx.modelPath	—	Model path to use when a model has no folder or file of its own.
llm.onnx.host	127.0.0.1	Host the server binds to.
llm.onnx.port	8090	Port the server binds to.
llm.onnx.readyTimeoutMs	120000	How long to wait (ms) for the server to become ready.