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

Public Integrations that do not require additional configuration are available by default in all Projects and do not require setup in the Integrations Dashboard.

The following Public Integrations do not require configuration.

ArXiv Research Integration

Searches for research papers on arXiv.org, an open-access repository covering fields such as physics, mathematics, computer science, and related disciplines.

  • com.globant.geai.arxiv_search: Searches for research papers on arXiv based on a query.

Image Generation Integration

Generates images from natural language descriptions using artificial intelligence models.

  • com.globant.geai.create_image: Creates an image based on a textual prompt and returns the URL of the generated image.

You can configure the model to use by defining the provider/model value. This configuration can be set:

  • In the Agent Guidelines, within the Agent Setup section.
  • Dynamically at runtime, allowing the consumer to specify the model.

If no model is specified, openai/gpt-image-1 is used by default.

The following models are supported:

  • openai/gpt-image-1
  • openai/dall-e-3
  • vertex_ai/imagen-3.0-generate-001
  • vertex_ai/imagen-3.0-fast-generate-001
  • vertex_ai/imagen-3.0-generate-002
  • xai/grok-2-image-1212

This Integration does not require manual configuration. However, in the Integration Parameters section, you must verify that the parameter is defined as follows:

  • Type: Context
  • Value: api_token

The value api_token is a reserved keyword used internally by the platform to resolve authentication at runtime and must not be modified.

PubMed Research Integration

Searches and retrieves biomedical research papers from PubMed.

  • com.globant.geai.pubmed_search: Searches for research papers in PubMed and returns a list of paper IDs (PMIDs) that match the query.
  • com.globant.geai.pubmed_fetch: Retrieves detailed information about research papers using a list of PMIDs, including title, authors, abstract, publication date, and journal information.

It is recommended to use both Tools together: first perform a search, and then retrieve detailed information using the returned PMIDs.

In the Tool Parameters section for com.globant.geai.pubmed_fetch, the parameters retmode and rettype are predefined.

Both parameters:

  • Have Type set to Application
  • Are marked as Required
  • Have their Value set at runtime

These parameters are part of the Integration definition and do not require manual configuration. Their values are resolved dynamically when the Tool is executed.

RAG Integration

Enables an Agent to interact directly with one or more RAG Assistants configured in the same Project.

  • com.globant.geai.chat_with_rag

For more information, see RAG Integration.

This Integration does not require manual configuration. However, in the Integration Parameters section, you must verify that the parameter is defined as follows:

  • Type: Context
  • Value: api_token

The value api_token is a reserved keyword used internally by the platform to resolve authentication at runtime and must not be modified.

Web Scraper Integration

Extracts structured information from websites.

  • com.globant.geai.web_scraper_httpx: Retrieves and processes content from a specified URL.

You must provide a URL as input. This Tool can also be used in conjunction with com.globant.geai.web_search to first discover relevant pages and then extract their content.

Last update: December 2025 | © GeneXus. All rights reserved. GeneXus Powered by Globant