Frequently asked questions

pixelseo.ai is an AI-powered image SEO tool. The primary feature — Optimize Images — analyzes images you already have and generates SEO-optimized file names, alt text, and JSON-LD schema markup for each one. There's also an AI image generator that produces website-ready images in 36 curated art styles with metadata included. It's built for SEO professionals, web designers, content marketers, ecommerce teams, and developers who want properly optimized images without the manual metadata work.

Optimize Images (at /process) is the core tool. Upload your own images — photos, product shots, screenshots, anything — and Pixel Vision (our AI pipeline) analyzes each image visually and generates an SEO-optimized file name, descriptive alt text, and ImageObject JSON-LD schema. You can optionally provide a page URL and the schema will include it as the content URL. Results are editable before download, and you can export as a CSV for bulk import.

Pixel Vision is the AI layer inside the Optimize Images tool. It uses Claude (Anthropic's vision model) to look at your actual image and generate metadata based on what's in it — not just a keyword you typed. This means the file name, alt text, and schema reflect the real image content, which produces more accurate and useful SEO output.

Formula Builder is an alternative mode within Optimize Images for when you already know what you want the file names to look like. Instead of AI-generated names, you provide a naming formula using a word bank — primary keywords, modifiers, descriptors — and the tool assembles structured file names at scale. Pixel Vision still writes the alt text and schema based on the actual image. It's useful for ecommerce catalogs, image libraries, and any situation where naming conventions are predetermined.

The Generate tool (at /generate) creates original AI images in 36 curated art styles and automatically produces the SEO metadata for each one. In Style mode, pick a style and optionally describe a subject. In Custom mode, write your own prompt from scratch. In Hybrid mode, start with a style prompt and edit it directly. All three modes return the same full SEO package.

There are 36 curated styles across two categories. Historical styles include Ancient Egyptian, Greek Classical, Byzantine, Renaissance, Baroque, Impressionism, Art Nouveau, Expressionism, Cubism, Surrealism, and more. Modern & Digital styles include Minimalism, Flat Design, Glassmorphism, Cyberpunk, Pixel Art, Photorealism, Cinematic, Cottagecore, Solarpunk, and others. You can browse all styles and see portrait and landscape examples before generating.

Yes. After processing or generating, all metadata fields are editable before you download. You can tailor file names, alt text, and schema to your specific page, target keyword, or CMS requirements. The CSV export includes all fields in a format ready to import.

For Optimize Images, you can upload JPG, PNG, WebP, and most common image formats. Output downloads are in WebP by default, which produces the smallest file sizes for web use. For the AI image generator, you can choose WebP, JPEG, or PNG output.

Images processed or generated through the tool are stored in your dashboard for 12 months. After that they are automatically removed. Download anything you want to keep permanently — the dashboard is a convenience, not a permanent archive.

Yes. The pixelseo.ai API exposes both core tools as REST endpoints. POST /api/v1/process-images/ accepts your images and returns SEO-optimized file names, alt text, and schema markup for each one — this is the batch pipeline. POST /api/v1/generate/ creates AI images and returns the full SEO package including the hosted image URL. Authentication uses an X-API-Key header. API keys are created and managed from your dashboard. Note the trailing slash on both paths — it's required.

POST /api/v1/process-images/ accepts multipart/form-data with your images, an optional site_url, optional page_urls per image, and optional provided_names for Formula Builder mode. It returns for each image: seo_filename, alt_text, description, JSON-LD schema, and token usage. Credits are only deducted for successfully processed images. Post to the trailing-slash path — the API redirects the non-slash form, and a multipart upload's body is dropped across that redirect.

POST /api/v1/generate/ returns: the image URL (hosted on our CDN), the SEO-optimized file name, alt text, JSON-LD schema markup, and your remaining credit balance.

Yes. Each image processed or generated — whether from the UI or the API — costs 1 credit. Credits are shared across both. There are no separate API-only plans.

Yes. n8n-nodes-pixelseo is a community node that wraps the batch API, so you can add image SEO to any n8n workflow without hand-building HTTP Request nodes. Install it from Settings → Community Nodes and search for n8n-nodes-pixelseo. Its main operation, Analyze Images (SEO Metadata), returns a filename, alt text, and ImageObject schema for each image, and takes a Page URL so the metadata reflects the page the image lives on. It also works as a tool for an AI agent.

Yes. The API is a standard REST API, so it works with any HTTP-capable automation tool — Make, Zapier, or your own code. A common pattern is pulling images from a CMS (WordPress or Shopify), sending them to /api/v1/process-images/, and writing the returned metadata back. For n8n specifically, use the dedicated community node above rather than raw HTTP requests.

There's a plugin for EmDash, the headless CMS for Astro sites on Cloudflare Workers — it connects your media library to pixelseo.ai so images land with the right filename, format, and alt text in one step. For other platforms, the REST API is the integration point.

pixelseo.ai uses a credit system — 1 credit per image, whether you're processing an existing image or generating a new one, and whether you're using the UI or the API. There are two ways to buy: one-time credit packs, which never expire; or the monthly Studio plan, which grants credits each month at a lower per-image rate. New accounts include 3 free credits so you can try it before paying. See the pricing page for current sizes and prices.

Packs for bursts, the monthly plan for a steady drip. If you're onboarding a client or backfilling an existing image library, buy a pack — those credits never expire, so you use them at your own pace and aren't paying for idle months. If you're adding new images every month across sites you maintain, the Studio plan gives a lower per-image rate with rollover, and you can cancel anytime.

Credits from one-time packs never expire — once purchased they stay in your account until used. Monthly Studio credits work as an allowance instead: unused credits roll over up to a cap rather than accumulating indefinitely. Either way, you're only charged for images that process successfully — a failed image costs nothing.

You'll receive a 402 response on the API or a prompt to purchase more in the UI. Your dashboard and all account features remain accessible — you just can't process or generate new images until you top up.

Search engines read file names as a signal for what an image contains. A file named 'impressionist-landscape-painting-homepage-hero.webp' tells Google exactly what the image is and how it relates to your page. A file named 'IMG_4823.jpg' tells it nothing. Descriptive, keyword-relevant file names are one of the simplest and most overlooked on-page SEO improvements you can make.

Good alt text describes what's in the image accurately and concisely, includes a relevant keyword naturally, and doesn't keyword-stuff. It's used by search engines to understand image content, by screen readers for accessibility, and as fallback text when images fail to load. It should read like a natural sentence describing what a sighted person would see.

JSON-LD schema is structured data added to your page's HTML that helps search engines understand your content in detail. For images, an ImageObject schema tells Google the image's name, description, URL, and context. This can improve how your pages appear in search results and makes your content eligible for rich results like Google Image Search enhancements.

Yes. Google indexes images based on what they contain and the metadata associated with them — not how they were created. AI-generated images with proper file names, alt text, and schema are indexed and can rank in Google Images just like any other image.

WebP is the recommended format for web use. It produces smaller file sizes than JPG or PNG at equivalent quality, which improves page load speed — a confirmed Google ranking factor. JPG is a good fallback for photos where WebP isn't supported. PNG is best when you need transparency (logos, icons). For most website images, WebP is the right choice.

Yes. Images you generate on pixelseo.ai are yours to use on your website, blog, or client projects. There are no licensing fees or attribution requirements.

Yes. Your generated images are stored privately in your account and are not publicly accessible or shared. The rename and convert tool processes files entirely in your browser — nothing is uploaded. We do not sell your data.

It depends on the tool. The Optimize Images pipeline (Pixel Vision) uses Anthropic's Claude vision model to analyze your images and generate the SEO metadata. The AI image generator uses Google's Imagen model via Vertex AI to create the images, with Claude handling the metadata side. Both tools use the same underlying output format.