Embeddings API (Python)

Methods, examples, and parameters for client.embeddings.

Module Overview

Generate vector embeddings for search, retrieval, and ranking pipelines.

Available Methods

client.embeddings.create() - Generate one or many embeddings.
client.embeddings.batch() - Submit large asynchronous embedding jobs.
client.embeddings.retrieve() - Get a previously created batch result.

Examples

Python

from therouter import TheRouter.ai

client = TheRouter.ai(
    api_key=os.getenv("THEROUTER_API_KEY"),
    base_url="https://api.therouter.ai/v1",
)

vectors = client.embeddings.create(
    model="openai/text-embedding-3-large",
    input=["routing", "guardrails"],
)

print(len(vectors.data[0].embedding))

embeddings-response.json

{
  "id": "req_01HXYZ123",
  "module": "embeddings",
  "status": "ok"
}

Parameter Types

Name	Type	Required	Description
model	string	Required	Embedding model identifier.
input	string \| string[]	Required	Text input or list of inputs.
dimensions	integer		Optional output dimension count.

SDK parity

Method signatures are aligned across SDKs so migration between TypeScript and Python stays predictable.