virtual staging

Furnish and style an empty (or badly furnished) room photo in a chosen interior style. The flagship tool. HD, watermark-free results. Costs credits from the use

liveunknown~5s typical

mcppixly

$0.0020

per call

plus $0.001 platform fee per run

Open MCP server; Cracked bills routing only

Provider Upstream docs

$Set up https://cracked-ai-pearl.vercel.app/SKILL.md, then use Cracked to run mcp-app-pixly-pixly /virtual_staging for me.

Input

Field	Type	Description
imageUrl	string	Public https URL of the source photo, or a data: URI. Either imageUrl or r2Path is require
r2Path	string	R2 object path from an upload ticket (POST /api/v1/uploads) — the alternative to imageUrl
style*	string	Style id from the Pixly styles catalog (lib/real-estate/styles.ts)
roomType	string	Room type hint; omit to let vision infer it from the photo
customInstructions	string	Optional extra instructions blended into the staging prompt
numImages	integer	Ignored — staging always returns one image. Call again for another take. default 1
stagingQuality	string (standard \| pro)	pro = Nano Banana Pro instead of Nano Banana 2, at 2 credits instead of 1. Same prompt and

Call it

curl

curl https://cracked-ai-pearl.vercel.app/v1/run \
  -H "Authorization: Bearer ck_live_..." -H "content-type: application/json" \
  -d '{"provider":"mcp-app-pixly-pixly","endpoint":"/virtual_staging","input":{"style":"...","numImages":1}}'

cli

npx cracked-ai run -p mcp-app-pixly-pixly -e /virtual_staging -i '{"style":"...","numImages":1}'

mcp

run_tool({ provider: "mcp-app-pixly-pixly", endpoint: "/virtual_staging", input: {"style":"...","numImages":1} })

Try it

Runs against your signed-in workspace balance. Sign in if you have not.