Embeddings API

Generate text embeddings through the Lightweight API gateway using the OpenAI-compatible endpoint.

Generate vector representations of text using the OpenAI-compatible embeddings endpoint. The same API key you use for chat and responses works for embeddings.

The embeddings endpoint uses the same API key and authentication as all other gateway endpoints. No additional configuration is required.

---

Endpoint

POST /v1/embeddings

Base URL: https://api.lightweight.one

Include an Authorization: Bearer YOUR_API_KEY header with every request.

---

Supported Models

Model	Default Dimensions	Configurable	Max Input Tokens
text-embedding-3-small	1536	1-1536 via dimensions	8,192
text-embedding-ada-002	1536	No	8,192

text-embedding-3-small is recommended for new projects. It supports configurable output dimensions for smaller, faster vectors. text-embedding-ada-002 is a legacy model.

---

Request Parameters

Parameter	Type	Required	Description
model	string	Yes	Model ID (see Supported Models above)
input	string or string[]	Yes	Text to embed. Single string or array of strings (max 2,048 items per batch). Also accepts number[] (pre-tokenized) and number[][] (batch pre-tokenized).
dimensions	integer	No	Output dimensions (only for text-embedding-3-small, range 1-1536)
encoding_format	string	No	"float" (default) or "base64"
user	string	No	End-user identifier for abuse monitoring

---

Examples

Single Input

curl -s -X POST https://api.lightweight.one/v1/embeddings \
  -H "Authorization: Bearer $LIGHTWEIGHT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "text-embedding-3-small",
    "input": "The quick brown fox jumps over the lazy dog"
  }'

---

Response Format

{
  "object": "list",
  "data": [
    {
      "object": "embedding",
      "embedding": [0.0023064255, -0.009327292, 0.015462338, "...1536 floats total"],
      "index": 0
    }
  ],
  "model": "text-embedding-3-small",
  "usage": {
    "prompt_tokens": 9,
    "total_tokens": 9
  }
}

Field	Description
object	Always "list"
data	Array of embedding objects
data[].object	Always "embedding"
data[].embedding	Float array (default) or base64 string (when encoding_format: "base64")
data[].index	Integer matching the input order
model	The model used
usage.prompt_tokens	Number of tokens in the input
usage.total_tokens	Same as prompt_tokens (embeddings have no output tokens)

---

Limits

• Maximum 8,192 tokens per input string
• Maximum 2,048 items per batch request
• Maximum 300,000 total tokens per request

---

Preview Environment

For testing, use the preview gateway at https://preview.api.lightweight.one. Replace the base URL in your requests to route through the preview environment.