
WordPress’s latest release has laid the foundation for built-in WordPress AI infrastructure, with connections to various AI models now natively integrated into the CMS.
With integrations from various AI providers, from OpenAI to Anthropic, WordPress now enables direct communication between your AI agent of choice and your website’s plugins and themes. This paves the way for generative content creation, task automation and general site optimisations – changes that sit alongside a broader shift in how AI is reshaping search results.
Table of Contents
What is WordPress’s new Connectors API?
The new Connectors API is a core framework for registering and connecting your WordPress website to external sources. Whilst this initially focuses on AI connections, it is the beginning of centralised API key management for the CMS. It provides plugin and theme developers with access to these tools to enhance their offerings. You can read the official WordPress Core developer notes on the Connectors API for a full technical breakdown.
Alongside this, WordPress.org has also published its free WordPress AI plugin, allowing your AI-powered tools to generate excerpts and content summaries, create featured images, and analyse your content for any key accessibility or readability errors.

What can the WordPress Connectors API actually do?
Features of the Connectors API:
Alongside the Connectors API comes a new UI experience – a centralised dashboard under Settings > Connectors, allowing users to connect and manage their AI integrations from one native interface.
WordPress have also released a series of development hooks allowing developers to call the site’s connected AI. This allows developers to use AI tools in the site’s plugins, whilst allowing site owners control over which AI platform they prefer.
Additional features include integration with the Abilities API, converting WordPress into a natively agentic platform, allowing plugins, themes, and the WordPress core to expose capabilities in human- and machine-readable formats, thus allowing AI agents to interact with all resulting WordPress features.
How should the WordPress Connectors API be used?
How can content editors use WordPress AI?
Websites perpetually publishing AI-generated content without human input have been hit hard by SEO algorithm updates due to a lack of contextual understanding. As we’ve explored before, just using AI without a clear strategy is rarely enough – and when it comes to your website content, the risks are very real. AI cannot replicate your brand voice or represent your products or services authentically and accurately without being part of your brand and team. Using these tools to fully write your website content can lead to issues that a thoughtful content creation strategy would avoid:
- Incorrectly optimised content, targeting keywords which aren’t relevant for your business.
- Obviously generated images undermining credibility and trust.
- High token/credit costs from frequent usage. (Whilst the Connectors API in itself is free, the tools you choose to connect to charge based on usage).
- Potentially inaccurate claims about your products and services
However, with proper use and human-led content, these new integrations within WordPress’s core open the door to powerful automation and a significant reduction in manual workload. These tools can be used to produce initial content drafts or generate contextual alt text for your media library in bulk. For those thinking about how this intersects with AI SEO, the implications for search visibility are significant. For content editors, this new workflow will enable you to generate meta descriptions and alt text at ease with contextual AI understanding.
How can developers use the WordPress Connectors API?
Developers will be able to utilise standard hooks to call the centralised connections and make use of AI platforms within their plugins or themes. And store owners can update their WooCommerce products, or generate discount tokens with a simple prompt.
Further to this, these new integrations will allow for dynamic translations, with providers translating content into various languages on the fly using advanced LLMs. The AI tools may also be used for suggesting relevant internal links from the existing database, or powering autonomous AI chatbots based on your site’s exact services, products or current inventory.
Is the WordPress Connectors API secure? Privacy risks explained
As with all development enhancements, security must be considered. WordPress 7.0 has introduced a number of security features to ensure these new AI integrations remain secure and compliant.
A new data classification system has been introduced, allowing database fields to be labelled as ‘public’, ‘sensitive’, or ‘personal’, strictly regulating what these AI tools are able to access on your site. However, correct usage of these labels relies on the effective adoption of this new system by third party plugin developers.
Further to this, a site owner adding their AI API key into the new centralised dashboard also adds this key, unencrypted, to the website database. Should the database be exposed, these keys become vulnerable. Therefore, WordPress 7.0 allows for direct hard-coded PHP constraints to define your API keys within the wp-config file instead. Ask your developers to action this for you for much safer and secure connections – or refer them to the WordPress developer documentation for the full technical guidance on implementing this securely.
In addition, site owners are now responsible for ensuring that their AI models are correctly configured to work on their websites. Consumer facing AI tools often use prompts to train future public models, meaning these tools could inadvertently expose your company’s proprietary data or leak your customer information into the public domain. It is important to ensure that your AI models are correctly configured to retain zero data, and disallowed from using your interactions for model training.
Despite these new security concerns, WordPress 7.0 actually solves previous security issues. Prior to this release, if you had an AI SEO plugin, an AI image plugin, and an AI chatbot plugin, you would not only have paid for three different subscriptions, and managed three different API keys, bloating the site’s codebase, but you would also have been relying on the security of these three individual integrations. With this unified API, your website can benefit from reduced plugin bloat, improved site loading speeds, and reduced third party developer security vulnerabilities.

What does wordpress AI mean for future development?
This new architecture is designed to reduce the fragmentation among current AI-enabled plugins, as many of these currently establish their own AI provider integrations and authentication systems. With WordPress 7.0, instead of each plugin team configuring their own AI provider integration, developers can access this unified, centralised AI connection and benefit from more streamlined, secure integrations.
The Connectors API framework also paves the way for future external services to be integrated with WordPress, making this one of the most significant updates we’ve seen to the WordPress platform in years. WordPress powers a significant portion of the web, and its approach to AI could influence how developers build AI-enabled publishing, e-commerce, and content-management tools in the years ahead.
WordPress is positioning itself to adapt as the AI landscape evolves. WordPress 7.0, dubbed “Armstrong,” officially moved the CMS into what many consider a more future-ready state, with native AI framework built directly into the platform – something you can explore in full in the official WordPress 7.0 release notes.
For developers, that means a more consistent way to add AI capabilities to WordPress projects while retaining the freedom to choose the models and services that best fit their needs.
For site owners, it’s worth understanding how these changes connect to broader shifts in SEO for AI search – and making sure your digital strategy is keeping pace.
If you’d like to talk through what this means for your site, our AI SEO team is here to help.

Get help with your WordPress website today and propel your business to online success.
Speak to Vicki – One of Our Web Development Experts.
WordPress AI Connectors FAQs
Q: Is the WordPress Connectors API free to use?
The Connectors API itself is built into WordPress 7.0 at no extra cost. However, the AI providers you connect to (such as OpenAI, Anthropic, or Google Gemini) charge based on usage, billing per API request. The plugin is free, but you’ll need to pay for API usage from your chosen AI provider – costs vary by provider and usage, though most offer free trial credits to get started.
Q: Which AI providers can I connect to in WordPress 7.0?
You can add an API key from your preferred AI provider – currently OpenAI, Google Gemini, or Anthropic Claude – via Settings > Connectors in your WordPress dashboard. Support for additional third-party providers such as Mistral and Cohere is expected in future releases.
Q: Will the WordPress AI Connectors API work with my existing plugins?
Not automatically. Most existing AI plugins won’t use the WordPress AI Client – they have their own API key settings built in and work independently. Only plugins built specifically to use the new shared API will benefit from the centralised connection. Adoption will grow as developers update their plugins to use the standard
Q: What can I actually do with WordPress AI once it’s connected?
The official AI Experiments plugin adds useful AI features to your editor and admin backend once a provider is connected, including title generation, excerpt generation, content summarisation, image generation, automatic alt text, and content suggestions for readability, grammar, accessibility, and SEO. Beyond that, developers can build plugins for tasks like chatbots, WooCommerce product descriptions, and internal link suggestions.
Q: Is my data safe when using the WordPress AI Connectors?
There are some important considerations. API keys entered via the Settings dashboard are stored unencrypted in the WordPress database by default, which is a vulnerability if your database is ever exposed. For better security, keys should be hard-coded into your wp-config.php file instead. You should also ensure your chosen AI provider is configured to retain zero data and is not using your interactions to train its public models, to protect both your business and your customers’ information.
Q: Could connecting AI to WordPress affect my SEO?
Yes, but how depends entirely on how you use it. AI can be a genuine help for tasks like generating meta descriptions, alt text, and content drafts – but sites that publish fully AI-generated content without human review have been penalised by Google’s algorithm updates. WordPress 7.0 ships the infrastructure, not the features – meaning the impact on your SEO is determined by how your team chooses to use these tools, not by the update itself.
Q: Could multiple plugins run up AI costs without me knowing
This is a real concern raised widely in the WordPress community. Once site owners connect a provider, multiple plugins can potentially tap into it – and the worry is that a plugin could make thousands of API calls before you notice the bill. The new centralised dashboard gives you one place to monitor usage across all plugins, which is a significant improvement over the previous fragmented setup. Checking which plugins have access to your connector and reviewing usage regularly is good practice.
Q: Is WordPress 7.0 AI GDPR compliant?
WordPress 7.0 introduces a data classification system that labels database fields as public, sensitive, or personal to regulate what AI can access. However, GDPR compliance ultimately depends on how your site is configured and which AI providers you use. Major US AI providers including OpenAI, Anthropic, and Google are enrolled in the Data Privacy Framework and offer Standard Contractual Clauses – but “available” is not the same as “in place.” You should confirm your provider’s data processing agreement and ensure consumer-facing AI tools are not using your data for model training.


