Embeddings

Endpoint

POST https://api.allroutes.ai/v1/embeddings

Authentication

Authorization: Bearer allroutes_sk_...
Content-Type: application/json

Request Body

Parameter	Type	Required	Description
`model`	string	Yes	Embedding model identifier (e.g., `text-embedding-3-small`, `text-embedding-3-large`)
`input`	string or array	Yes	Text to embed. A single string or an array of strings for batch embedding.
`encoding_format`	string	No	Output format: `"float"` (default) or `"base64"`
`dimensions`	integer	No	Truncate embeddings to this many dimensions (supported models only)

Request Examples

Single Input

curl https://api.allroutes.ai/v1/embeddings \
  -H "Authorization: Bearer allroutes_sk_..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "text-embedding-3-small",
    "input": "The quick brown fox jumps over the lazy dog"
  }'

Batch Input

curl https://api.allroutes.ai/v1/embeddings \
  -H "Authorization: Bearer allroutes_sk_..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "text-embedding-3-small",
    "input": [
      "First sentence to embed",
      "Second sentence to embed",
      "Third sentence to embed"
    ]
  }'

Response Format

{
  "object": "list",
  "data": [
    {
      "object": "embedding",
      "index": 0,
      "embedding": [0.0023064255, -0.009327292, 0.015797347, ...]
    }
  ],
  "model": "text-embedding-3-small",
  "usage": {
    "prompt_tokens": 9,
    "total_tokens": 9
  }
}

Response Fields

Field	Type	Description
`data`	array	Array of embedding objects
`data[].embedding`	array of floats	The embedding vector
`data[].index`	integer	Position of the input in the request array
`model`	string	Model used to generate the embeddings
`usage.prompt_tokens`	integer	Number of tokens consumed
`usage.total_tokens`	integer	Total tokens (same as prompt_tokens for embeddings)

SDK Examples

Python

from allroutes import AllRoutesClient

client = AllRoutesClient(api_key="allroutes_sk_...")

# Single input
resp = client.create_embedding(
    model="text-embedding-3-small",
    input="The quick brown fox jumps over the lazy dog",
)
vector = resp["data"][0]["embedding"]
print(f"Dimensions: {len(vector)}")

# Batch input
resp = client.create_embedding(
    model="text-embedding-3-small",
    input=["First sentence", "Second sentence", "Third sentence"],
)
for item in resp["data"]:
    print(f"Index {item['index']}: {len(item['embedding'])} dims")

Go

embResp, err := client.CreateEmbedding(ctx, &allroutes.EmbeddingRequest{
    Model: "text-embedding-3-small",
    Input: "The quick brown fox jumps over the lazy dog",
})
if err != nil {
    log.Fatal(err)
}

fmt.Printf("Dimensions: %d\n", len(embResp.Data[0].Embedding))
fmt.Printf("Tokens used: %d\n", embResp.Usage.TotalTokens)

Supported Models

AllRoutes routes embedding requests to the appropriate provider based on the model name:

Model	Provider	Dimensions	Max Input Tokens
`text-embedding-3-small`	OpenAI	1536	8191
`text-embedding-3-large`	OpenAI	3072	8191
`text-embedding-ada-002`	OpenAI	1536	8191
`embed-english-v3.0`	Cohere	1024	512
`embed-multilingual-v3.0`	Cohere	1024	512

Use the Models API to list all available embedding models with current pricing.

Error Codes

Status	Description
`400`	Invalid request (e.g., empty input, unsupported model)
`401`	Invalid or missing API key
`402`	Insufficient credits
`413`	Input exceeds the model's maximum token count
`429`	Rate limit exceeded
`500`	Provider error (automatically retried)

Cost Optimization

Embeddings are typically cheap, but large batch jobs add up. Three high-impact tactics:

Batch where possible -- a single request with a 100-element input array beats 100 sequential requests on both latency and the per-request fixed overhead.
Use smaller dimensions -- text-embedding-3-small is sufficient for most retrieval workloads at one-fifth the cost of text-embedding-3-large.
Truncate dimensions -- pass dimensions: 512 on supported models to shrink vectors and downstream storage cost without retraining.

See Cost Optimization for benchmark numbers.