Llama Nemotron Embed VL 1B V2 (free) API: toman pricing and code

curl https://api.uttapen.ir/v1/embeddings \
  -H "Authorization: Bearer sk-up-…" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "nvidia/llama-nemotron-embed-vl-1b-v2:free",
    "input": ["How long is the electric kettle warranty?", "Warranty period for kitchen appliances"]
  }'

from openai import OpenAI
import numpy as np

client = OpenAI(base_url="https://api.uttapen.ir/v1", api_key="sk-up-…")

docs = ["The electric kettle warranty is 18 months.", "Orders ship within 48 hours."]
question = "How long is the kettle guaranteed for?"

vecs = client.embeddings.create(model="nvidia/llama-nemotron-embed-vl-1b-v2:free", input=docs + [question]).data
M = np.array([v.embedding for v in vecs])
M /= np.linalg.norm(M, axis=1, keepdims=True)
scores = M[:-1] @ M[-1]
print(docs[int(scores.argmax())])

import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://api.uttapen.ir/v1", apiKey: "sk-up-…" });

const docs = ["The electric kettle warranty is 18 months.", "Orders ship within 48 hours."];
const { data } = await client.embeddings.create({ model: "nvidia/llama-nemotron-embed-vl-1b-v2:free", input: [...docs, "How long is the kettle guaranteed for?"] });

const norm = (v) => { const n = Math.hypot(...v); return v.map((x) => x / n); };
const vecs = data.map((d) => norm(d.embedding));
const q = vecs.at(-1);
const dot = (a, b) => a.reduce((s, x, i) => s + x * b[i], 0);
const best = vecs.slice(0, -1).map((v, i) => [dot(v, q), i]).sort((a, b) => b[0] - a[0])[0][1];
console.log(docs[best]);