Convert text into dense vector representations for search, clustering, and semantic similarity.
The embeddings endpoint converts text into high-dimensional floating-point vectors. Use them to build semantic search, recommendation systems, document clustering, or any pipeline that needs to compare meaning rather than exact strings.
The request and response shape are compatible with the OpenAI embeddings API, so any SDK or library targeting that spec will work without changes.
| Field | Type | Required | Description |
|---|---|---|---|
model | string | Yes | Embedding model slug, e.g. openai/text-embedding-3-large |
input | string or array | Yes | Text to embed. Pass a string for a single input or an array of strings for batch embedding. |
encoding_format | string | No | "float" (default) or "base64" |
dimensions | integer | No | Truncate output vectors to this many dimensions (model-dependent) |
Results appear in the same order as your input array. The usage.prompt_tokens count covers all inputs combined.
Embeddings are billed at the model's input token rate in DA, deducted from your balance the same way as chat completions. The X-Balance-Available header reports your remaining balance after each request.
Filter the model catalog for embedding-capable models at /models?capability=embeddings or via the API: