Embeddings

POST /v1/embeddings returns one vector per input text. The embedding model is baai/bge-m3 (alias bge-m3): multilingual, up to 8,192 input tokens per text, 1,024-dimension vectors. Its price is on the model page.

curl

curl https://api.thetoken.dev/v1/embeddings \
  -H "Authorization: Bearer $TOKEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "baai/bge-m3",
    "input": ["Where is my order?", "أين طلبي؟"]
  }'

Python

import os
from openai import OpenAI

client = OpenAI(base_url="https://api.thetoken.dev/v1", api_key=os.environ["TOKEN_API_KEY"])

result = client.embeddings.create(
    model="baai/bge-m3",
    input=["Where is my order?", "أين طلبي؟"],
)
vectors = [item.embedding for item in result.data]
print(len(vectors), len(vectors[0]))  # 2 1024

TypeScript

import OpenAI from 'openai'

const client = new OpenAI({ baseURL: 'https://api.thetoken.dev/v1', apiKey: process.env.TOKEN_API_KEY })

const result = await client.embeddings.create({
  model: 'baai/bge-m3',
  input: ['Where is my order?', 'أين طلبي؟'],
})
console.log(result.data.length, result.data[0].embedding.length) // 2 1024

The response (vectors shortened):

{
  "object": "list",
  "model": "baai/bge-m3",
  "data": [
    { "object": "embedding", "index": 0, "embedding": [-0.0312, 0.0187, 0.0054] },
    { "object": "embedding", "index": 1, "embedding": [-0.0298, 0.0201, 0.0049] }
  ],
  "usage": { "prompt_tokens": 12, "total_tokens": 12 }
}

Good to know

• input is a string or a list of strings; results come back in the same order (index).
• Compare vectors with cosine similarity. Texts in different languages with the same meaning land close together, as in the example.
• Only input tokens are billed; embeddings have no output price.
• Some clients send token numbers instead of text (LangChain’s Python OpenAIEmbeddings by default); send text. See the LangChain guide.