AI Entity Intake Alpha Introduces Reviewer-Controlled Entity Extraction for Drupal

Editorial teams using AI-assisted workflows in Drupal now have a new reviewer-controlled intake option with the release of AI Entity Intake 1.0.0-alpha1. The contributed module, developed by Drupal contributor Jesse L., who goes by the handle dobe on Drupal.org, is designed to convert unstructured text such as meeting notes, emails, catalogue specifications, and product information into draft Drupal entities for review before publication. The release supports Drupal 10.5 and Drupal 11.2.

The module focuses on keeping publishing decisions under human control. Instead of automatically creating or publishing content, AI Entity Intake presents candidate entities that reviewers can approve, reject, or edit manually. According to the project page, every entity remains in draft form until explicitly approved.

The alpha release supports multiple entity types through separate integration submodules, including taxonomy terms, media items, profile entities, commerce products, paragraphs, and group relationships. Site administrators can install only the integrations needed for a specific workflow.

The module also supports compound entity generation. The project description states that the extraction workflow can propose a landing page with nested paragraph components or a commerce product with variations during a single intake session. A recursive processing pipeline then prepares those related entities for review and approval.

Additional features include duplicate detection using exact, Levenshtein, and token-set matching strategies, optional asynchronous extraction for longer-running AI tasks, and bulk approval or rejection controls for larger review queues. Sites using the Group module can also attach approved entities to existing group contexts automatically.

AI Entity Intake depends on the AI module and requires a configured provider that exposes structured-output chat capabilities. The project documentation lists OpenAI and Anthropic among supported providers. Additional requirements include the JSON Field and Key modules.

The maintainers position the module as distinct from autonomous AI workflow systems. The project page compares AI Entity Intake with AI Agents, which are designed for more automated execution flows, while AI Entity Intake keeps approval decisions with human reviewers. The documentation also references AI Content Suggestions as a separate approach focused on modifying existing entities rather than creating new ones from raw text.

The project remains in alpha development and is not covered by Drupal’s security advisory policy. The project page warns that users should evaluate the software carefully before production use.

Installation is available through Composer using composer require 'drupal/ai_entity_intake:^1.0@alpha'. The current development branch, 1.0.x-dev, was updated on 20 May 2026 at 19:38 UTC.