Actions14
- Product Management Actions
- Product Management > Consumptions Actions
- Results Actions
- Usage > V1 > Images Actions
- Usage > V1 > Models Actions
- Usage > V1 > Open AI Actions
- Usage > V2 > Open AI Actions
Usage > V1 > Open AI → Create Embeddings
AI-generatedSummary
Create embeddings by posting to the Infomaniak AI Tools API endpoint /1/ai/{product_id}/openai/v1/embeddings with the required parameters: product ID (from Product Id), input data (from Input, which can be a string, array of strings, array of integers, or array of arrays of integers), and model (from Model). Optionally specify encoding format (base64 or float), mode (index or query), and task description (for query mode) via Additional Body Fields. Authenticates via OAuth2 (default) or API Key. By default, outputs each embedding object from the response's data array as separate workflow items; if Return Full Response is enabled, outputs the entire API response as a single item.
Inputs
- Authentication (required) — Choose between OAuth2 (default) or API Key for authentication.
- Product Id (required) — The identifier for your AI Tools product (obtain via GET /1/ai).
- Input (required) — The data to embed. Can be a string, array of strings, array of integers (tokenized sentence), or array of arrays of integers (batch of tokenized sentences).
- Model (required) — The model name to use for embedding.
- Additional Body Fields — Optional collection to specify encoding format (base64 or float), mode (index or query), and task description (for query mode).
- Return Full Response — If set to true, returns the full API response; otherwise, returns only the embedding data items (default: false).
Output shape
By default, returns multiple items (one per embedding) each containing the embedding object (with embedding vector, index, etc.) in the json field. If Return Full Response is true, returns a single item containing the full API response in the json field.
The node does not support pagination for this operation (pagination: none).
Examples
Example 1: Embedding a single sentence
Set Product Id to your product ID, Input to the sentence text, Model to the desired embedding model (e.g., 'bge-large-en-v1.5'), and leave Additional Body Fields and Return Full Response at defaults.