Hugging Face models from a Node.js function

main.js is a RunR Node.js function (XaaS) that calls two bound models with the global fetch: embeddings through the alias EMBED and a chat answer through CHAT. Set up the models and bindings as in the Python example, paste main.js into the editor and deploy.

curl -s -X POST https://<function>.<your function domain>/ -d '{"text": "hello"}'
curl -s -X POST https://<function>.<your function domain>/ -d '{"action": "chat", "text": "What is RunR?"}'

The requests use the OpenAI-compatible routes (/v1/embeddings, /v1/chat/completions) that TEI, vLLM, TGI and RunR's hf-serve answer, so the OpenAI SDK for Node.js works as well:

const OpenAI = require('openai');
const client = new OpenAI({ baseURL: process.env.RUNR_MODEL_CHAT_OPENAI_BASE, apiKey: 'unused' });
const reply = await client.chat.completions.create({
  model: process.env.RUNR_MODEL_CHAT_MODEL,
  messages: [{ role: 'user', content: 'hello' }],
});

(add openai to a FaaS project's package.json). See docs/huggingface.md for every variable RunR sets per binding.