Create embeddings

Creates embedding vectors for the given input with the selected embedding model and debits the cost from your organization's embedding credits.

POST /v1/embeddings

Turn model input into embedding vectorsThis request consumes the organization's embedding credits.

Choose an available modelCheck its input types, dimensions and price.
Supply text or imagesSend the input items to the selected model.
Match vectors to inputsUse each returned index to recover input order.

Before you start

Use your organization API key on the control-plane origin.

Connection setup

Choose an available, priced model from the embedding catalog. Requests consume embedding credits.

Request body

application/json

Limits and accepted values

model must be a catalog id that is currently available and priced (otherwise 400 for an unknown or unavailable model, 403 for an unpriced one);
a missing or non-string model is rejected with 422.
input is declared optional for parsing but the route rejects an absent, null or empty-array input with 400: send a single item (a string, or an object with exactly one of text or image) or an array of items, up to the deployment's batch limit (400 above it).
Any additional fields (for example input_type, query or document) are forwarded unchanged to the embedding service.

Field	Type	Required
`model`	`string`	Required
`input`	`any JSON`	Required

Read the result

Each item in data contains an embedding and its input index. Match that index to your original input order. dim is the vector length; usage.total_tokens reports usage for the billed request.

Response statuses

HTTP status reference

Status	What to do
200	OK
400	Correct the request before retrying.
401	Check the credential and Authorization header.
402	Check feature access or available credits.
403	Check tenant access, roles and permissions.
415	Send the required Content-Type.
422	Correct the values or unsupported combination.
500	Check the outcome before retrying a write.
502	Check service availability and the write outcome.
503	Honor Retry-After when present; check write outcomes.

Request example

curl --include --request POST "https://api.originchain.ai/v1/embeddings" \
  --header "Authorization: Bearer ${OC_ORG_API_KEY}" \
  --header "Content-Type: application/json" \
  --data-binary '{
  "model": "<AVAILABLE_MODEL_ID>",
  "input": [
    "Trail mug with insulated steel"
  ]
}'

Response example

Examples use synthetic values to show the response format.

Response status

{
  "object": "list",
  "model": "<selected-model-id>",
  "dim": 3,
  "data": [
    {
      "object": "embedding",
      "index": 0,
      "embedding": [
        0.1,
        0.2,
        0.3
      ]
    }
  ],
  "usage": {
    "total_tokens": 6
  }
}

Shape illustration using a synthetic three-value vector; actual model dimensions, embedding values and billed tokens come from the selected model.

Download the OpenAPI specification for the complete response schemas.

Before you startCopy Anchor Link

Request bodyCopy Anchor Link

Read the resultCopy Anchor Link

Response statusesCopy Anchor Link

Before you start

Request body

Read the result

Response statuses