---
title: "Text embeddings"
description: "Convert text into vectors for semantic search and retrieval."
lastUpdated: 2026-10-22
---

Send `POST https://infistar.ai/v1/embeddings` with `Authorization: Bearer YOUR_API_KEY` and JSON content. The Azure-compatible path is `/v1/engines/{model}/embeddings`.

| Field | Required | Meaning |
| --- | --- | --- |
| `model` | Yes | Exact available embedding model ID |
| `input` | Yes | A string or array of strings; batch limits depend on the model |
| `encoding_format` | No | Vector format, normally `float`; `base64` when supported |
| `dimensions` | No | Output dimensions, only for models that support adjustment |

Check model availability with your key and review [current pricing](https://infistar.ai/pricing). Do not mix vectors from different embedding models in one similarity index.

For document preparation, retrieval and reranking, see [RAG integration](/en/integration-guides/rag-retrieval-quickstart).