AI Featured Image Generator

Yes, for at least one provider. All free-plan providers have free tiers or pay-as-you-go pricing: OpenAI, Stability AI, Grok/xAI, Google Gemini (Pro only), or Meta Muse (Pro only).

The plugin is free; you pay your provider per image:

• OpenAI: varies by model and generation quality (Low is the cheapest); see OpenAI pricing
• Stability AI: Stable Image Core is roughly ~$0.003/image; other models vary; see Stability AI pricing
• Grok/xAI: $0.02/image (Grok Imagine Image), $0.04/image (Grok Imagine Image 2.0) or $0.05/image (Grok Imagine Image Quality)
• Google Gemini (Pro): Nano Banana 2.1, Gemini 3.1 Flash and 3 Pro vary by resolution
• Meta Muse (Pro): $0.01/image

Yes. Select your provider and enter its API key in the plugin settings. The page shows only the options relevant to your selection. Free users can switch between OpenAI, Stability AI, and Grok/xAI; Pro adds Google Gemini and Meta Muse.

Only your API plan's quota with your chosen provider. The plugin generates one image per request. The free plan has one lifetime cap: 20 images requested by AI assistants (MCP). Pro removes it. Everything else, including Bulk Generate, has no plugin limit.

OpenAI (Free & Pro): GPT Image 2 (default, with strong text rendering and visual reasoning). Pro adds GPT Image 2.6 Flare (fast and recommended, with sharper detail) and GPT Image 2.6 Sunburst (the most detailed), both with Extra High and Max generation quality. Stability AI: Stable Image Core (free): fast, high quality, best value. Pro adds Stable Image Ultra (best quality), SD 3.5 Large / Large Turbo / Medium / Flash. Grok / xAI: Grok Imagine Image (free, $0.02/image), Grok Imagine Image 2.0 (Pro, $0.04/image) and Grok Imagine Image Quality (Pro, $0.05/image). Google Gemini (Pro only): Nano Banana 2.1 (1K/2K/4K, the default), Nano Banana 2 (Gemini 3.1 Flash, 1K/2K/4K, deprecated by Google) and Nano Banana Pro (Gemini 3 Pro, best quality, 1K/2K/4K). All support 10 aspect ratios. Meta Muse (Pro only, beta): Muse Image 1.0, Meta's in-house reasoning image model. New and not yet widely tested, so please report anything odd. 9 aspect ratios, up to 1600px, $0.01 per image.

21 built-in styles that work with all providers: None, Realistic, Artistic, Cartoon, Minimalist, Sketch, Watercolor, 3D Render, Pixel Art, Cyberpunk, Fantasy, Anime, and Technicolor, plus the Trending group: Toy Figure, Miniature World, Claymation, Cinematic, Lo-Fi Cozy, Vintage Instant Photo, Magazine Collage, and Paper Cut. With Stability AI you also get a Stability Presets group of 16 native API-level presets: Photographic, Cinematic, Anime, Digital Art, Comic Book, Fantasy Art, 3D Model, Analog Film, Enhance, Isometric, Line Art, Low Poly, Neon Punk, Origami, Pixel Art, and Tile Texture. Pro users can also create unlimited custom styles (see below).

Yes. Go to AI Images → Settings → Brand & Styles → Custom Styles, click "Add Style", and enter a name plus a prompt description (e.g. "in a dreamy impressionist style"). Custom styles appear in all style dropdowns and work exactly like built-in styles.

One global set of brand rules that keeps every AI image on-brand. Go to AI Images → Settings → Brand & Styles and save up to 3 brand colors, a style description, a tone, and avoid-rules; optionally add a reference image (used by Google Gemini to match your visual style) and a logo watermark with position (four corners or centre), size, opacity, margin and horizontal/vertical offset controls, with a live preview you can drag to place the logo. The kit is applied automatically wherever you generate: post editor, Bulk Generate, the Posts list, and the AI Image Generator block. The post editor, Bulk Generate, and the block each have a "Use Brand Kit" checkbox you can uncheck for a one-off image; the Posts list always applies the kit.

• OpenAI: 7 sizes with GPT Image 2 and 2.6: 1024x1024 (1:1), 1024x1536 (portrait), 1536x1024 (landscape), 1792x1024 (16:9), 1024x1792 (9:16), 1920x640 (3:1), and 640x1920 (1:3)
• Stability AI: 9 ratios (1:1, 16:9, 21:9, 2:3, 3:2, 4:5, 5:4, 9:16, 9:21)
• Grok: 9 ratios (1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 2:1, 1:2)
• Gemini (Pro): 10 ratios (1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9): 1K/2K/4K on Nano Banana 2.1, Gemini 3.1 Flash and 3 Pro
• Meta Muse (Pro): 9 ratios (1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 2:1, 1:2)

Yes, with a Stability AI API key. Once a featured image is set, a "Stability AI Tools" panel appears in the post editor with three modes: Conservative (faithful 4x), Creative (AI-enhanced 4K), and Fast (quick 4x). Conservative and Creative process asynchronously and may take up to ~3 minutes; the result automatically becomes your new featured image.

Yes, from the same "Stability AI Tools" panel. Click "Remove Background" and Stability AI isolates the subject and saves a transparent PNG as your featured image, which suits product shots and portraits.

Text that tells Stability AI what not to include, e.g. "blurry, watermark, text, low quality". Set defaults in AI Images → Settings → Post Images. Negative prompts have no effect with OpenAI, Gemini, or Grok.

• OpenAI: WebP (recommended), PNG, or JPEG
• Stability AI: PNG free; Pro adds JPEG and WebP
• Grok: JPEG free; Pro adds WebP (converted after generation)
• Gemini (Pro): WebP (default, converted after generation) or PNG
• Meta Muse (Pro): WebP (default), PNG, or JPEG, returned directly by the API

WebP conversion for Gemini and Grok needs a WebP-capable image library on your server (GD with WebP, or ImageMagick with the WebP delegate). If it's missing, the plugin safely falls back to the provider's native format and shows a notice with the reason. Generation never fails. Ask your host to enable WebP if needed. OpenAI WebP is unaffected because the API returns it directly.

• OpenAI: Output Compression slider (1–100, default 90) plus a Low/Medium/High generation quality tier (Extra High and Max on GPT Image 2.6, Pro)
• Stability AI: quality is inherent to the model you choose
• Grok: 1K or 2K resolution
• Gemini (Pro): 1K/2K/4K on Nano Banana 2.1, 3.1 Flash and 3 Pro
• Inline images (all providers): Automatic Quality Optimization picks quality and resolution based on the image's display size in your content

Yes. Add the "AI Image Generator" block from the Media category, enter a prompt, and generate images anywhere in your content. The block supports the same providers, styles, and aspect ratios as featured images; configure defaults in the Inline Images settings tab.

Every AI image (featured, inline, or bulk) automatically gets alt text in the media library. Configure it under AI Images → Settings → General → Alt Text: choose the post title (default), the generation prompt, or a custom template with {title}, {prompt}, {style}, and {site_name} placeholders. e.g. {title} - {style} style - {site_name}. By default, manually edited alt text is preserved on regeneration; enable "Overwrite Alt Text on Regeneration" to always refresh it.

Click "Generate Prompt from Content" in the post editor and AI writes an image prompt from your post title and body. It's saved to the post and pre-filled in the Custom Prompt field so you can review or edit before generating. The same option exists in Bulk Generate, creating a unique prompt per post. Text model per provider: gpt-6-luna (OpenAI), grok-4.7 (Grok), Gemini 3.6 Flash (Gemini, Pro), Muse Spark 1.3 (Meta Muse, Pro). Stability AI has no text API, so the button shows a notice when it's the active provider.

Yes. Go to AI Images → Settings → AI Assistants and follow the 3 steps: install the free MCP Adapter plugin (one click), create a key (one click), and connect your app with the ready-made config. Claude Desktop, Cursor, Claude Code and VS Code are covered with step-by-step instructions, and the tab ticks step 3 by itself the first time your assistant calls the site, which costs nothing. Then start each request with "On" and your site name, for example "On My Site, make a featured image for my latest post". That one rule works in every app and keeps several sites apart. The assistant can only do what your WordPress user can do, and only the plugin's image and audio tools are exposed. Free sites get 20 assistant images plus finding, auditing and alt text; Pro is unlimited and adds the tools below. On a Free site the Pro tools still show up, but answer with an upgrade link instead of running. Needs WordPress 6.9 or newer. claude.ai in the browser and ChatGPT cannot connect yet because they need an OAuth sign-in that the MCP Adapter does not provide.

Free:

• Make images: a featured image for a post (with a prompt, style, shape, provider or model of your choice) or any image for the Media Library. 20 images in total.
• Find: posts, pages and any other post type by title, author, category, tag, date, or whether they have a featured image; read a post to write a better prompt; find images in the Media Library by name.
• Image audits: count and list images by format, shape, pixel size, file size, upload date, missing alt text, blur or cut-off edges, across posts of any type or the whole Media Library; measure one image for sharpness and cut-off edges; look at a preview to judge the picture itself.
• Image text: change an image's title, alt text, caption or description.

• Image tools: rename an image (every post keeps pointing at the right file), convert JPG or PNG to WebP (every post switches to the lighter copy), upscale, or remove the background.
• Audio: make or refresh a post's narrated recording.
• Place images: set a Media Library image as a featured image, or insert it into a post.
• Brand: read and change the Brand Kit (colours, tone, description, avoid list, reference image), pick and place the logo, preview the kit, stamp the logo onto the featured images posts already have (no AI call) and undo it, find posts whose image predates the current kit, and set the kit up from the site name, tagline, theme colours and logo.
• Products (WooCommerce): make store images from a product's photo, one image or a main image plus gallery set. The assistant shows you the result first and only puts it on the product when you say so. It can also check every product image against your style without making anything.
• Animated and vector: make an image as an animated GIF, or a Media Library image as an SVG drawing (GPT Image 2.6).

Yes. Go to AI Images → Settings → Audio, choose ElevenLabs (the most natural voices), Grok (the lowest price, with your existing Grok key) or Google Gemini (with your existing Gemini key and its free tier), pick a voice and play a sample. Then open a post and click Generate audio in the AI Featured Image box (on a product, in the AI Product Images box). The recording is saved in the Media Library and a "Listen to this post" player appears above the content. The box shows the estimated cost first, and tells you when the post changed since the recording so you can regenerate it. You can hide the player on any post, or turn all players off on the Audio tab.

Go to AI Images → Image Audit and click Check images. It counts the images in your posts, or in the whole Media Library, that have no alt text, are over 600 KB, are still JPG or PNG instead of WebP, or look blurry or cut off. Free sites see the counts and a short preview of each list. Pro lists every image and fixes them: Fill missing alt text writes alt text for all of them at once using your Alt Text settings on the General tab (the post title, the image prompt or your own template, with no AI service involved), Convert to WebP makes a smaller copy and points your posts at it, and you can edit alt text, titles and captions in the list.

Yes. Turn on Auto-generate on Publish under AI Images → Settings → Post Images. When a post is published without a featured image, one is made in the background a few seconds later with your current provider, model, style, size and Brand Kit, so publishing is never slowed down. You can choose whether the prompt is written from the post content (one extra text request) or taken from the title and style. The Posts list shows when it is generating, when it is done, and the reason if it failed. It runs through WP-Cron; if your host turns WP-Cron off, make sure a real cron job runs it.

Yes. On the Audio tab, turn on Podcast feed, add a show title and a square cover image (1400 to 3000 pixels), and save. The tab then shows your feed address, usually yoursite.com/feed/podcast/. Submit that address once to Spotify for Podcasters, Apple Podcasts Connect, YouTube Music, Amazon Music or any other podcast app. The apps check the address on their own and add each new recording as an episode, usually within an hour. Nothing is sent from your site: the apps read the feed like any visitor. Recordings hidden on their post, drafts and password-protected posts are left out. SoundCloud cannot read a feed address, so it is not supported.

Yes. Set the style once on AI Images → Settings → Product Images (background, type of shot, mood, your Brand Kit logo, model, size and the images in a product set). For lifestyle shots the AI picks a setting that suits the product; to ask for a particular one, type it in the product box, for example "on a beach towel at sunset". Then open a product: the AI Product Images box uses the product's own photo, so the product stays exactly as it is and only the background, light and scene change. Click Generate image or Generate product set, look at the results, and choose Set as main image, Add to gallery or Replace. Nothing changes until you click one. Restore originals puts back the images the product had before. To do the whole shop, open Bulk images on the same Product Images tab. Products without a photo cannot be picked. Products never get the AI Featured Image box, so there is one place for posts and pages and one for products.

GPT Image 2.6 Flare and Sunburst make transparent images themselves; AIFIG switches GPT Image 2 to 2.6 Flare for these. Gemini (Nano Banana) cannot make transparent images: when it is asked to, it often paints a grey checkerboard that then shows on your shop. So with Gemini, AIFIG makes the image on white and your Stability key cuts the background out, and the word transparent is never sent to Gemini. Without a Stability key it uses GPT Image 2.6 Flare if you have an OpenAI key. Every transparent product image is also checked on your server, and a drawn checkerboard is never saved.

On AI Images → Settings → Product Images → Bulk images, choose Match photos to your style (and, if you like, Which photos: all, main images only or gallery photos only), then click Load Products. Each photo's background colour, shape, size and, for transparent backgrounds, file type are checked against your style on your server, for free, and only products with photos that differ are listed, with the exact number of photos. Each of those photos is remade from that same photo and put back in its place, so a close-up stays a close-up. Main images follow your Main image choice; gallery photos become Studio shots. For Studio and 3D the type of shot is not checked, so to turn every main image into a 3D shot, use Main image only. With a Showroom or Lifestyle main image, every main image not already made in that shot is remade. Lifestyle and showroom photos made with the plugin on purpose are kept as they are. The original photos are always kept: Restore originals in a product's AI Product Images box puts them back.

Yes. Go to AI Images → Settings → General → Availability, check the post types you want, and save. The generator panel then appears in the editor for those post types. Free users are limited to Posts and Pages.