Transformers.js / ONNX Setup Guide

DevDiff supports running open-source models in-process using @xenova/transformers (ONNX runtime). This provider enables zero-dependency, local AI execution directly within Node.js worker threads without an external Ollama daemon.

---

⚙️ Configuration Setup

.devdiff.config.js Setup

export default {
  ai: {
    providers: [
      {
        name: "transformers-local",
        url: "transformers://Xenova/Qwen1.5-0.5B-Chat",
        quantization: "q4",
        priority: 1,
      },
    ],
  },
};