POST /api/v1/route/embeddings: OpenAI-Compatible Embeddings - Belvedir

embeddings.create against the router base URL runs OpenAI’s embedding models on Belvedir’s account with your bv_live_ key. The same client you use for chat completions works unchanged.

const router = new OpenAI({
  baseURL: "https://platform.belvedir.ai/api/v1/route",
  apiKey: process.env.BELVEDIR_API_KEY,
});
const res = await router.embeddings.create({
  model: "text-embedding-3-small",
  input: ["first document", "second document"],
});

Request

Authorization: Bearer token with your API key, or use the x-api-key header.
Body: a standard OpenAI embeddings body (model, input, and the optional dimensions and encoding_format). Bodies cap at 4 MB (413 beyond it).

Models

Any text-embedding-* id is accepted, with or without the openai/ prefix, but an unpriced one is refused with 400. The priced ids are text-embedding-3-small, text-embedding-3-large, and text-embedding-ada-002. Embeddings are model-pinned: there is nothing to route, and "auto" is not accepted.

Cost

Input tokens are metered like every routed call, at the rates on the Pricing page (Inference → Pricing), attributed to the calling API key. The cost of the call rides the x-belvedir-cost response header.

Limits

The 402 spend gate and the rate limit (25 requests per second per API key, burst 300, adjustable per organization) are the same as on chat completions. The full table is on Capabilities and Limits.