Description
From a prompt or a plain draft to a finished, well-formatted post, in your theme’s style.
AI writes text well. The problem is the wall of text it leaves in WordPress. Woof Creator lays your content out as a complete page or post: headings, sections, lists, callouts, FAQs, buttons and real photos, as native blocks you can still edit by hand. Then keep talking to it: “make the intro shorter”, “add a key takeaways box”, “use our brand color #0a7”.
Woof Creator is in beta. It works, and we use it on real sites, but AI design is new territory and you may hit rough edges. If something looks wrong or doesn’t work, please tell us with the “Send feedback” link in the plugin (or at https://woofcreator.com/contact-us/): we read every message and fix what you report.
Key features
- Prompt to formatted post. Describe a page or post, paste text you already have, or attach the document it’s in (Word, Markdown or text), and get it laid out in sections with the right block for each part. Refine it by chatting and undo any change.
- Looks like your theme. Uses your theme’s colors, fonts and spacing so new content matches the rest of your site, or creates a custom style from your description.
- Native blocks, no lock-in. Output is standard block editor (Gutenberg) blocks; the Classic editor and Elementor are supported too. Edit by hand in the block editor or Elementor and the AI keeps your changes. Your content is saved as ordinary blocks, so it stays in your pages even without the plugin.
- Edit with AI in the block editor. A sidebar for the whole page and an “Edit with AI” button on the block toolbar for a single section.
- Blocks you already have. The AI can use blocks from your other plugins, such as pricing tables and forms, and WordPress blocks such as social icons, embeds and latest posts.
- Real photos, not placeholders. Openly licensed photos from Openverse (no key needed), saved in your Media Library with their credits.
- AI-generated images. Have images made from a description with your own OpenAI or Google Gemini key, or the AI connectors in WordPress 7.0+, when you ask for them or for every new image.
- Bring your own AI. Use your own account with a major AI provider or a local model, or the AI connectors in WordPress 7.0+ with no key at all. Your key is stored encrypted, requests go straight from your site to your provider, and you see the token use of every reply.
You can also rebuild a page from a screenshot of a design you like: attach it to a message and the AI recreates the layout with your content and theme, then compares the result with your screenshot and fixes what differs.
How it works
- Connect an AI provider under Woof Creator → Settings (or pick “WordPress AI” on WordPress 7.0+).
- Open Woof Creator → Designer and describe your page or post, or paste your draft.
- Refine it by chatting, then save it as a draft and publish it from the editor.
Supported AI providers
OpenAI, Anthropic Claude, Google Gemini, OpenRouter, Mistral, Groq, DeepSeek, xAI (Grok), local models through Ollama or any OpenAI-compatible endpoint, and the AI connectors set up in WordPress 7.0+ (Settings → Connectors).
Pro add-on
A separate, optional Pro add-on (see https://cloud.woofcreator.com/pro) adds Pexels and Pixabay photos, SEO titles and descriptions with FAQ structured data, and pages from PDF, OpenDocument, RTF and web page files, a link to a web page or a Google Doc. Everything in this plugin works without it.
Documentation: https://woofcreator.com/docs/
Source code
The JavaScript in build/ is compiled from the unminified source in the src/ folder, which is included in this plugin. To rebuild it, run npm install and then npm run build in the plugin folder. The compiled code includes the modern-screenshot library (MIT license, https://github.com/qq15725/modern-screenshot).
Trademarks
The code is GPL; the brand is not. “Woof Creator” and its logo are trademarks of Competitors App SRL. You are welcome to fork the plugin under the GPL, but a modified version must use its own name and branding, not ours. Unmodified copies may keep the name. For brand questions or to report misuse, see https://woofcreator.com/.
External services
This plugin sends your design requests, together with your site name and tagline (and, to OpenRouter, your site’s address), your theme’s style settings and the current page draft, to the AI provider you configure, using your own API key. Images you attach to a chat message are uploaded to your Media Library; their URLs are sent to the provider and, if “Send images to the model” is on for that provider, a scaled-down copy of each image as well. Documents you attach to a chat message are read in your browser: their text is sent to the provider with that message and kept with the conversation on your site, and images inside a Word document are uploaded to your Media Library like attached images. When “Send images to the model” is on for the provider, a screenshot of the page preview is also sent with a message if you turn on “Attach a screenshot of the page” in the chat, when your message is about how the page looks, or when the AI asks to look at the page; it is not saved on your site. After the AI rebuilds a page from a reference image, it compares a screenshot of the result with the reference and fixes what differs: up to 2 more requests, set under Settings → “Check a rebuilt page against the reference” (0 turns it off). When a step of an answer doesn’t work out, such as a photo search that finds nothing, the AI may try again: up to 2 more requests, set under Settings → “Retry when a step fails”. Clicking “Suggest ideas for this site” in the chat sends your site’s name, tagline, address and locale, the titles of up to 30 published pages and 15 published posts and your most used categories to the provider, to suggest ideas for new pages and posts (never drafts, private or password-protected content); the ideas are kept on your site for a week or until it changes, and “Suggest page and post ideas” in Settings hides the button. Opening Settings with a provider set up asks the provider for its list of models, which also checks your key. Nothing else is sent until you send a message.
If you choose “WordPress AI” as the provider, the same data is sent through the WordPress AI Client to the AI service of the connector you set up under Settings → Connectors, under that service’s terms.
- OpenAI: https://openai.com/policies/terms-of-use, https://openai.com/policies/privacy-policy
- Anthropic: https://www.anthropic.com/legal/consumer-terms, https://www.anthropic.com/legal/privacy
- Google Gemini: https://ai.google.dev/gemini-api/terms, https://policies.google.com/privacy
- OpenRouter: https://openrouter.ai/terms, https://openrouter.ai/privacy
- Mistral: https://mistral.ai/terms#terms-of-service, https://mistral.ai/terms#privacy-policy
- Groq: https://groq.com/terms-of-use, https://groq.com/privacy-policy
- DeepSeek: https://platform.deepseek.com/downloads/DeepSeek%20Open%20Platform%20Terms%20of%20Service.html, https://cdn.deepseek.com/policies/en-US/deepseek-privacy-policy.html
- xAI: https://x.ai/legal/terms-of-service, https://x.ai/legal/privacy-policy
- Google Fonts (only if you turn on “Allow Google Fonts”, which is off by default, and a custom style uses web fonts: visitors’ browsers then load the fonts from Google): https://policies.google.com/terms, https://policies.google.com/privacy
Generated images (Settings → “Photos and images”, used only with a key or connector and only when you ask for generated images, unless you choose “For every new image”): the description of the image is sent to the image service you choose, OpenAI or Google Gemini (terms and privacy policies above), using your own API key, or through the WordPress AI Client to your connector’s AI service. The image that comes back is saved in your Media Library. It’s charged to your account with that service.
Photos (Settings → “Photos and images”): when the AI adds a photo to a page, or you search in the image picker, the search words (a few English words describing the photo) are sent to Openverse, and the chosen photo is downloaded into your Media Library from the site Openverse lists it on (such as Flickr or Wikimedia Commons). Openverse needs no key; you can turn stock photos off instead. Search answers are kept on your site for a day. Your browser loads the picker’s thumbnails from Openverse or those sites.
- Openverse: https://docs.openverse.org/terms_of_service.html, https://wordpress.org/about/privacy/
Videos: a video element with a YouTube or Vimeo link, only when you or the AI add one, embeds the player on the page, so visitors’ browsers load it from that service.
- YouTube: https://www.youtube.com/t/terms, https://policies.google.com/privacy
- Vimeo: https://vimeo.com/terms, https://vimeo.com/privacy
An embed block (for example of a post on X, Instagram or Spotify), when you or the AI add one, is WordPress’s own Embed block: WordPress fetches the embed from that service, and visitors’ browsers load it, under that service’s terms. The author box shows the author’s Gravatar picture, as WordPress does in comments (Gravatar: https://wordpress.com/tos/, https://automattic.com/privacy/).
Only when you submit the “Send feedback” form, the plugin sends your message to our service (cloud.woofcreator.com, run by Competitors App SRL), with the details the form lists before you send it: site URL, plugin, Pro add-on, WordPress and PHP versions, theme, language, the AI provider and model you chose, the Elementor version and whether the site is a multisite. Your email address is sent only if you ask for an answer. No page content, conversations or API keys are sent. Terms and privacy: https://competitors.app/terms, https://competitors.app/privacy
Installation
- Install Woof Creator from Plugins → Add New Plugin and activate it.
- Go to Woof Creator → Settings, choose an AI provider, paste your API key and click “Save and test connection”. On WordPress 7.0 or newer you can instead choose “WordPress AI” to use the connectors set up under Settings → Connectors, with no key in the plugin.
- Go to Woof Creator → Designer, click “+ New page” or “+ New post” and describe what you want, or paste text you already have.
FAQ
What does it cost?
The plugin is free. You pay your AI provider directly for what you use, at their prices; Settings shows your usage and an estimated cost. Local models through Ollama cost nothing.
Is my content private?
Nothing goes to an AI provider until you set one up. After that, opening Settings checks your key by listing the provider’s models, “Suggest ideas” in the chat asks it for ideas that fit your site when you click it, and your requests and page drafts go straight from your site to the AI provider you chose, with your own key. We never receive your content or your API key. See “External services” below for the full list.
Does it work with my theme?
It reads your theme’s colors, fonts and spacing and builds with them, so it works with block themes and classic themes. Settings → Theme report shows what it knows about your theme and how well it matches.
Will it change my words?
When you paste a draft or attach the document it’s in (.docx, .md or .txt) and ask for it to be laid out, it keeps your text and adds structure: headings, lists, callouts and sections, following your document’s own headings, and places the document’s images where they were. It rewrites only when you ask it to.
Which AI model should I use?
Larger models give better layouts. If results look plain or incomplete, try your provider’s strongest model first. Each provider’s default model is a good starting point.
Can I edit the result by hand?
Yes. The result is normal blocks (or Classic editor HTML, or Elementor widgets). Edit them as usual; when you ask the AI for another change, it keeps your edits.
What does Pro add?
See “Pro add-on” above. Pro is a separate plugin with extra features; everything in this plugin works without it.
Can it update my old posts?
Not yet. Keeping existing posts up to date is planned for a later version.
Screenshots





Changelog
0.24.2
- New: “Improve with AI” in the chat turns a few words, or the files you attached, into three detailed requests to pick from.
- Improved: tighter spacing between sections. Classic themes no longer add their own padding inside every section, and only the first and last sections of a page keep large padding.
0.24.1
- Fixed: “Match my theme” previews now use your theme’s own stylesheets and Customizer settings, so the preview shows the same fonts and colors as your site (GeneratePress, Astra, Kadence and other classic themes).
- Improved: a photo search that Openverse answers with a temporary server error is tried once more.
0.24.0
- First release on WordPress.org.