Skip to main content

Embeddings

Generate vector representations of text for search, retrieval, recommendation, clustering, and other tasks.

Request​

POST /embeddings

Request Body
{
  "model": "text-embedding-3-small",
  "input": ["Hello", "Welcome to Qiyiguo AI, a unified API for multiple model providers."]
}

Response​

Response Example
{
  "data": [
    { "index": 0, "embedding": [0.06, 0.03, 0.08] },
    { "index": 1, "embedding": [0.09, 0.10, 0.20] }
  ],
  "model": "text-embedding-3-small",
  "usage": {
    "prompt_tokens": 6,
    "total_tokens": 6
  }
}

Best Practices​

  • Use consistent tokenization and text normalization strategies
  • Choose a vector store that supports cosine or dot-product similarity
  • Manage embedding versions and dimensions to avoid mixing them in production