AI HomeDesign MCP

Enhance a photo: better lighting, sharpness and colour balance. No style needed. Provide image_url OR image_base64. area: indoor | outdoor (detect from the image). options (multi, optional): add-fire-to-fireplace, add-screen-to-tv.

Fetch an order, including its input assets and any processes run on it.

Fetch a process: its `status` (pending/processing/done/failed) and, when done, the `final_assets` with downloadable result image URLs. If a webhook result has already arrived for this process, it is returned immediately.

Fetch a single project by id.

Fully redesign a FURNISHED room (walls, floors, furniture, decor) in a style. Same workflow as virtual_staging: never pick the style yourself — ask the user; detect the room type from the image. Provide image_url OR image_base64.

List every AI HomeDesign tool this MCP can run, grouped by category. Returns each tool's slug, display name, description and whether it needs one or two input images. Start here to discover what is available.

List your projects (paginated). Optional free-text `search` on address.

List the allowed design styles, room types and widget options. With no argument, returns the global style/room/area lists. With a tool_slug, returns that tool's widgets and their allowed values. The staging/design helpers ONLY accept these values — anything else is rejected before calling the API.

Auto-detect and remove furniture/clutter from a room photo (decluttering), leaving a clean empty space. No style needed. Provide image_url OR image_base64.

Start an AI job on an existing order (full control over every tool). `asset_map` maps each of the tool's input-slot keys to an asset id from the order. `widgets` is the raw list of {slug, item_slugs}. Use `describe_tool` to get the exact slugs. Returns `process_ids` to poll with `get_process`.

Replace the furniture in an ALREADY-FURNISHED room with a new design style. Same workflow as virtual_staging: ask the user for the style, detect the room type from the image. Provide image_url OR image_base64.

Furnish an EMPTY room photo with AI furniture in a chosen design style. Provide the photo as `image_url` OR `image_base64` (raw base64 or a data: URL). WORKFLOW — follow exactly: • Do NOT choose a `style` yourself. Call this first with style empty; it returns the list of design styles. Present them to the user and let THEM pick, then call again with their chosen style. • Detect `space` (room type) from the image yourself and pass it (bedroom, kitchen, living-room, ...). Only if you truly cannot …

Poll a process until it is `done` or `failed` (or until timeout), then return the final result image URLs. Respects the API's 5s minimum poll interval.