CMS & AI
A CMS that writes with you — and talks to your AI.
Pages, categories, products and blog in one admin. Thirteen AI pipelines write the boring 80 %. And an MCP endpoint lets Claude do the rest — previewing every change before it lands.
store_update_product_batch
betadry_run: true · 3 items · 0 rows written
Pellet stove Comfort 12 kW
meta_description
— 12 kW pellet stove with automatic ignition, 92% efficiency and a 7-day programmer. Free delivery in Bulgaria.
Pellet boiler Pro 25
meta_description
— 25 kW pellet boiler for homes up to 250 m² — self-cleaning burner, 5-year warranty, installation available.
Pellets, 15 kg bag
meta_description
— ENplus A1 wood pellets, 15 kg — low ash, high heat. Pallet prices and delivery across Bulgaria.
Preview first. Apply when you're happy.
applyContent is where stores stall.
Two hundred products, each needing a description, a meta description, an FAQ and structured data — nobody writes that by hand, so it stays empty, and empty pages do not rank. Most CMSs give you a text box. Yours should give you a writer.
Edit everything in one place
- Pages, hierarchical categories with navigation placement, products and blog — one admin panel, pre-compiled, no build step for you
- Rich editor: images, video, multi-column layouts, collapsible sections
- Image gallery with thumbnails; an Open Graph image generator for every product, page and post — your brand overlay and title rendered onto the photo
- Meta title and description on every entity; publish and draft; the static site rebuilds only what changed
- Composable homepage: seven built-in sections — hero, featured products with four sources and four sort orders, category grid, promo banner, latest posts, newsletter, contact — reorderable per store; custom sections and component overrides that survive platform updates
AI writes the boring 80 %
Thirteen ready pipelines, run in batch through our AI proxy with spend caps:
- Fix long description · short description · meta description · rewrite long description
- Google Shopping description · Schema.org JSON-LD · Markdown for AI agents · FAQ section
- Optimize page content for SEO and SEM · keyword research · responsive-search-ad analysis · image prompts
- Bulk tools for prices, descriptions, markdown, content fixes and MSRP import
- A content auditor scores every product against e-commerce best practice and lists what is missing — thin copy, identifiers, images
- A prompt manager with variables: set the voice once, every pipeline reuses it
Bring your own AI — MCP built in
Every store exposes an MCP endpoint. Connect claude.ai, Claude Desktop or Claude Code and say: “Write meta descriptions for every product that has none — preview first.”
- Twelve tools: store overview · list pages and categories · get page · search products · get product · catalogue listing · Search Console performance · Merchant feed status · update product · update 25 products at once · update page or category · request rebuild
- Writes are dry-run by default and return a per-field before/after diff — nothing is saved until you apply
- GTINs are check-digit validated; slugs and prices are deliberately not writable; every call is audit-logged
- Token authentication with throttling — your AI sees only your store
Readable by AI crawlers too
- llms.txt and llms-full.txt generated automatically for products, pages and posts
- Structured data on every page, so assistants and search engines read the same facts
Under the hood
- PHP 8.3 API + MariaDB
- Lexical editor — content stored as JSON + HTML
- Open Graph pipeline in GD → WebP
- MCP: stateless Streamable HTTP JSON-RPC
- Rebuild markers, static export on demand
See the editor and the AI pipelines on your catalogue
A 20-minute demo on your own products is worth more than a feature list.