Drupal is embracing AI as if they were always meant to be together. As Dries puts it, “Drupal accidentally prepared for AI”. What made Drupal 8 painful to adopt is what’s really powering its AI abilities today.

Along with its core capabilities, Drupal continues to offer just the right tools to step up your game. 

Drupal AI modules are contributed modules that connect your Drupal site to AI providers like OpenAI, Anthropic, and Google Gemini - enabling capabilities like content generation, automated alt text, semantic search, on-page translation, and much more - without custom code.

In this blog, I shall walk you through eight Drupal AI modules worth actually installing in 2026 - what each one does, who it fits, and where it still has rough edges. 

Key Takeaways:

• The core AI module is the foundation: install it first, configure one provider, then layer on the rest.
• AI Search replaces keyword matching with intent-aware, vector-based results.
• AI Image Alt Text automatically generates accessible, SEO-ready alt text for every image.
• SmartLinker AI (built by Specbee) automates internal linking directly inside CKEditor 5.
• AI Agents is the most capable submodule in the ecosystem, and also the least ready for careless production use.

1. What does the Drupal AI (Artificial Intelligence) module do?

The Drupal AI module is a unified abstraction layer that connects Drupal to 48+ AI providers - OpenAI, Anthropic, Google Gemini, Mistral, and self-hosted models via Ollama - through a single consistent API. Install this one first. It allows you to configure, swap, and run different models for different tasks without writing custom per-provider code. 

The best part is that you can configure GPT-4 for content generation and use a cheaper model for embeddings. You can swap providers without rewriting module code, run different models for different operation types, such as text, image, speech, translation - all from one admin screen.

Without this layer, every AI integration would need custom per-provider code. This module solves that problem once.

• Best for: Any team getting started with Drupal AI.
• Requires: Drupal 10.4+, PHP 8.1+, Key module (pulled in automatically via Composer).

2. How does AI Search improve site search in Drupal?

Traditional keyword search matches words. AI Search matches intent.

It's a submodule of the core AI module that integrates with the Search API to store content as vector embeddings (mathematical representations of content meaning that allow similarity-based retrieval rather than keyword matching) in databases like Milvus, Pinecone, or Qdrant (purpose-built vector databases designed to store and query embeddings at scale). When a user searches for "why my images load slowly on mobile," the module doesn't look for those exact words. It retrieves content that matches the underlying meaning.

The approach is Retrieval-Augmented Generation (RAG) - a method where relevant content is retrieved from a vector database and passed to an LLM alongside the user's query, so the model generates answers grounded in your actual site content rather than its training data. For content-heavy sites in higher education, government, publishing, or healthcare, this kind of search meaningfully narrows the gap between what users ask and what they actually find.

Its setup is not lightweight. You'll need a running vector database and solid familiarity with the Search API module. But for the right site, the payoff is real.

• Best for: Enterprises with large, structured content libraries.
• Rough edge: Requires external infrastructure (vector DB, embeddings configuration) before anything else works.

3. Can AI generate alt text for images automatically in Drupal?

Yes, and the AI Image Alt Text module handles it well.

It's part of the core AI module ecosystem and generates accessible alt text for images using AI vision models. The process keeps a human in the loop, allowing editors to review and adjust before saving, and supports multilingual alt text based on entity language.

For teams with large media libraries, this saves meaningful time. Consistent, descriptive alt text improves both WCAG accessibility compliance and image search rankings. The module integrates directly with Drupal's image widgets, so it fits existing editorial workflows.

There's also a companion submodule - AI Image Bulk Alt Text - for backfilling alt text across images already in the library.

• Best for: Accessibility-focused teams, multilingual sites, and media-heavy content operations.

4. What does the AI CKEditor submodule add to the editing experience?

The AI CKEditor submodule puts an AI assistant directly inside CKEditor 5, letting editors rewrite, translate, summarize, or adjust the tone of selected text without leaving the editing interface. 

For content teams that live in the editor, this removes a lot of friction. No copy-pasting from an external chatbot. No context switching. The interaction stays where the work is.

The module supports custom prompts, so organizations can configure standard instructions like "rewrite in plain language," "translate to Spanish," "summarize to three sentences" that editors can trigger with a click.

• Best for: Content editors who spend most of their time in CKEditor 5.
• Part of: Core AI module (no separate download required)

5. How does AI Translate work for multilingual Drupal sites?

AI Translate provides one-click AI-powered translation for multilingual Drupal sites, creating actual translated content nodes in the database, which means proper multilingual SEO rather than a browser-side overlay.

For organizations running content in multiple languages, this cuts the manual loop of exporting strings, sending them out for translation, and reimporting. The translation happens in place and produces properly translated content entities.

However, it’s worth flagging that machine translation quality varies by language pair. This module works best as a first-pass drafting tool, particularly for high-volume content where a human reviewer picks up the AI draft.

• Best for: Organizations managing multilingual content at scale.
• Part of: Core AI module ecosystem.

6. What are AI Automators in Drupal?

AI Automators set rules so that AI performs field-level tasks automatically: populate fields, summarize content, extract text from uploaded files via OCR, generate transcripts, or scrape external URLs.

The distinction from the editor-facing submodules is that AI Automators are rule-based triggers that run AI tasks automatically - populating fields, summarizing content, extracting text via OCR, or generating transcripts, without requiring a human prompt each time. 

According to Drupal.org, Automators can chain prompts together in as simple or complex a workflow as needed - making them the practical starting point for meaningful AI applications inside Drupal. The supported operation types cover LLM text generation, web scraping, OCR file extraction, and audio transcription, with further extension available through ECA (Entity, Content, Action) integration, a Drupal module for building event-driven automation rules.

• Best for: Sites with high content volumes and predictable field-population patterns.
• Rough edge: Automated field population at scale needs clear data governance and quality monitoring.
• Part of: Core AI module ecosystem.

7. How does SmartLinker AI handle internal linking in Drupal?

Internal linking is genuinely tedious at scale. Every time a new article goes live, someone has to scan through older content, find relevant keywords, and add links by hand. At volume, that process is inconsistent and easy to deprioritize. SmartLinker AI, developed by Specbee, automates internal linking inside CKEditor 5 by scanning content for relevant phrases and suggesting links in real time, without editors having to do it manually. 

The module uses GPT-4 integration and vector-based search through Milvus and Zilliz Cloud to scan content, detect key phrases, and suggest links directly inside CKEditor 5. Editors can review suggestions, set keyword targeting rules, choose which content types to link to, and control the number of links per piece.

The module also validates links in real time. If a linked page is moved or renamed, SmartLinker AI updates the reference site-wide, preventing broken links and 404 errors.

At DrupalCon Atlanta 2025, SmartLinker AI was reviewed by core Drupal AI team members including James Abrahams of Freely Give Ltd., one of the primary contributors to the AI module ecosystem. Abrahams engaged with the module directly and noted the specificity of the use case - automated internal linking being a gap the core ecosystem doesn't currently address. 

Its setup requires the core AI module, Milvus or Zilliz Cloud for vector search, and a few supporting modules. It's not a one-click install. Once it's running, the editorial workflow is straightforward.

• Best for: Content teams publishing high volumes of articles, case studies, or documentation.

8. What can AI Agents do in Drupal - and where does the risk come in?

AI Agents is the most capable submodule in the ecosystem. It enables Drupal to take autonomous actions, such as creating views, updating configuration, managing content, running multi-step workflows, based on a prompt or scheduled trigger, without human intervention at each step.

It allows Drupal to take autonomous actions based on a prompt or scheduled trigger: create views, update configuration, manage content, assign taxonomy, or run multi-step workflows. These aren't assistants that wait for a human to click something. They act on their own.

The module uses a plugin-based architecture, so developers can build custom agents for specific site needs. It integrates with the AI module's Chatbot and Assistants API. Current maintainers include Marcus Johansson, James Abrahams, and Andrew Belcher.

The right setup is a controlled, well-monitored workflow with clear scope limits. Teams that have already worked with Automators and AI CKEditor are better positioned to use Agents safely.

• Best for: Experienced Drupal developers building bounded, well-tested automation workflows.
• Rough edge: Experimental status. Production use requires careful scoping and active monitoring.
• Part of: Core AI module ecosystem.

Drupal AI modules vs. the WordPress alternative

WordPress users working with AI typically rely on Jetpack AI, Yoast AI, or standalone plugins - each covering a narrow capability with little interoperability. Drupal's approach is different. The core AI module acts as a shared infrastructure layer, so every submodule (CKEditor, Alt Text, Translate, Automators, Agents) draws on the same provider configuration and API abstraction.

For teams that need more than one AI capability, that architecture matters. Adding a second or third Drupal AI feature doesn't mean configuring a second or third provider connection from scratch.

The primary difference you should look at is that WordPress AI plugins are faster to install. On the contrary, Drupal's ecosystem requires more upfront setup but rewards it with significantly more flexibility at scale.

How do these modules stick together?

They are a stack, not a menu. Here's how the layers fit together: 

A reasonable starting point: install the core AI module, connect one provider, and test prompts in AI Explorer. AI Image Alt Text and AI CKEditor are the lowest-friction entry points from there. AI Search and AI Agents are better suited for teams with more technical infrastructure already in place.

Final thoughts

The Drupal AI ecosystem has grown fast. What started as a handful of experimental integrations is now a structured, funded, production-ready framework. According to Drupal.org, nearly 14,000 sites report using the core AI module as of mid-2026. Dries Buytaert's published 2026 roadmap commits 28 partner organizations and over 50 full-time contributors to eight core AI capabilities.

The foundation of Drupal's AI module ecosystem is stable. The submodules cover real editorial and technical needs, and the roadmap has serious organizational backing.

Where to start depends on your situation:

• Content editors dealing with repetitive tasks - AI CKEditor or AI Image Alt Text. Low setup friction, immediate editorial value.
• Sites with large multilingual content volumes - AI Translate. First-pass drafts at scale, human review for quality.
• Enterprises with structured content libraries - AI Search. Requires vector database infrastructure, but meaningfully improves findability.
• Teams publishing high volumes of articles - SmartLinker AI. Automates the internal linking work that almost always gets deprioritized.
• Experienced developers building automation workflows - AI Automators, then AI Agents. Start bounded. Monitor carefully.

If you're evaluating where Drupal AI fits into your roadmap, our Drupal AI services team has hands-on experience with the modules above.

Frequently Asked Questions

How much does running Drupal AI modules cost?

The modules themselves are free and open source. The cost comes from AI provider API usage - OpenAI, Anthropic, and Google Gemini all charge per token or per request. For high-volume use cases like Automators or AI Translate, those costs can add up quickly. Self-hosted options via Ollama eliminate API costs but require server infrastructure. Build your use case on a test environment and measure API consumption before committing to production volume. 

Can SmartLinker AI be used on non-Drupal sites?

No. SmartLinker AI is built specifically for Drupal and depends on CKEditor 5, the Drupal AI module, and a compatible vector database (Milvus or Zilliz Cloud). It is not a standalone tool and cannot be ported to WordPress or other CMSes. 

How do I monitor AI-generated content quality on a live Drupal site?

There is no native quality-monitoring dashboard in the current ecosystem. Teams using Automators or AI Agents at scale should establish editorial sampling workflows, reviewing a percentage of AI-generated field values on a regular cadence. For high-stakes content (legal, medical, compliance), treat AI output as a draft that requires human sign-off before publishing. 

What's the setup complexity across these modules?

It varies considerably. AI CKEditor and AI Image Alt Text are low friction once the core module is configured. AI Search and SmartLinker AI require external infrastructure - vector databases, embedding configurations, API keys - which adds meaningful setup overhead. AI Agents should only be attempted by teams comfortable with Drupal configuration management and the risks of agentic automation.

Is SmartLinker AI the same as the Drupal AI module?

No. SmartLinker AI is a separate contributed module built by Specbee that uses the Drupal AI module as part of its stack. The core AI module is a community project. SmartLinker AI is a Specbee product.

Can these modules work with self-hosted or privacy-first AI setups? 

For most use cases, yes. The core AI module supports Ollama for self-hosted models. This matters for organizations with data residency requirements or GDPR obligations. Provider compatibility varies across submodules and operation types, so it's worth checking the specific combination you need before committing to a setup.

Do these modules work with Drupal 11?

The core AI module 1.3.x supports Drupal 10.5+ and 11.2+. Most submodules follow the same compatibility range. Check each project page on Drupal.org for specifics.