Table of contents
Official Content
  • This documentation is valid for:

The Corpus detail view lets you edit a specific Corpus. To open it, go to the Corpus Dashboard (Build section of the Agentic Hub) and click on Edit button on that Corpus, available in both Card view and Table view. This view is organized into four tabs: Configuration, Embeddings, Chunking, and Ingestion. The second tab, titled Embeddings, lets you define how the Corpus generates the vector embeddings used to index its Resources and to search them by semantic similarity.

Embeddings

The Embeddings tab displays the following fields:

  • Provider: Selects the service provider used to generate the embeddings for the Corpus, for example openai, azure, awsbedrock, cohere, or gemini. This selection is mandatory.
  • Model Name: Selects the embeddings model. The available options depend on the selected Provider. For example, with the openai Provider you can choose text-embedding-3-small, text-embedding-3-large, or text-embedding-ada-002. For the full list of supported models for each Provider, see Supported Embeddings Models.
  • Dimensions: Sets the number of dimensions of the generated embeddings. The default depends on the selected model, for example 1536 for OpenAI models.
  • Type: Defines the processing type used by the embeddings model. This applies only to certain models. Check the model documentation before setting it.
  • Use Proxy: Routes the embeddings requests through the Globant Enterprise AI proxy. When enabled, all requests are logged. It is enabled by default.
  • Cache: Stores the generated embeddings for the current Organization and Project to avoid recalculating them. It requires Use Proxy to be enabled.
  • Batch Size: Sets how many items are sent together in each embeddings request during ingestion. The default is 100. You may need to adjust it depending on the selected model.
  • Mode: Selects the embeddings mode. Leave it empty for text embeddings, or select multimodal for image embeddings when the model supports it.

Note that changing the Provider, Model Name, or Dimensions after the Corpus has been ingested is recommended to be followed by a reingestion of its Resources, so the index stays consistent.

A Save button is fixed in the bottom-right corner of the view. Click on Save to persist your changes. After a successful update, a Changes saved indicator is shown in the bottom-left corner.

Availability

Since version 2026-07.

Last update: 2026 | © Globant S.A. All rights reserved.