The exact-fit search is what a keyword catalog cannot do — 'a shelf for a 425mm gap' is a question about numbers. get_related_items and coordinate_storage then close the usual second problem: the shelf arrives and the boxes do not fit it.
A hosted product server over a curated furniture and home catalog. Search is dimension-aware down to the millimetre, and results carry the matching storage, protection and accessory items rather than a single product.
- search_products and get_product_detail — keyword, dimension, price, colour, category and brand search, then inner dimensions, consumables and compatible storage
- suggest_by_space — give a space in millimetres and get what fits, rotation aware
- identify_product — a visual description in, product candidates with model numbers out
- coordinate_storage, compare_products and get_related_items — shelf plus box sets with quantities and cost, side-by-side comparison, and accessory chains
- find_replacement — a discontinued model resolved to successors, ranked by dimension compatibility
- calc_room_layout — rectangle packing with placement coordinates
- get_curated_sets, get_popular_products and list_categories
- search_rakuten_products and search_amazon_products for live listings
- diagnose_ai_visibility — an audit of a site's llms.txt, robots.txt, JSON-LD and OGP
- Three prompt workflows: room_coordinator, moving_checklist and product_showdown
Nothing to install — point a client at https://ai-furniture-hub.onrender.com/mcp over Streamable HTTP, with an SSE endpoint also available. It can also be run locally with npx or from a clone; API keys are optional there and it falls back to mock data without them.
One command — npx ai-furniture-hub