MeltFlex AI — Claude Code MCP server: what it does, how to install
Turns a room photo into a photorealistic interior, facade or garden redesign, and builds 3D furniture models or 3D worlds from the same photo.
claude mcp add meltflex -e MELTFLEX_API_KEY=<MELTFLEX_API_KEY> -- npx -y meltflex-mcpWhat MeltFlex AI does
MeltFlex AI talks to the public meltflexai.com API and gives an agent tools for reworking a space: changing a room's style, placing furniture, swapping floors, walls, stairs, doors or windows, redoing a kitchen or bathroom, or restyling a garden and facade. One mode turns a rough 3D draft into a photo, another edits only a masked region of the image. Beyond stills, the server can generate walkthrough videos, convert a furniture photo into a GLB model, turn a 2D floorplan into a 3D model or rendered picture, and turn a room photo into an explorable 3D world.
Calls go to /api/v1/generate or related endpoints with parameters for mode, resolution and quality level; the three 3D builds can return an HTTP 202 with a poll URL instead of an immediate result. Running any of this needs a personal MeltFlex API key, and that key only gets issued once there's an active paid subscription on the MeltFlex site, so a free account alone won't be enough.
The server runs through npx -y meltflex-mcp with MELTFLEX_API_KEY set in the environment. Image generations spend account credits and get refunded automatically if a job fails.
What it helps with
- restyle a room photo into a different interior look
- drop a specific furniture piece into a room photo at the right scale and color
- swap floors, walls, doors or windows on an existing photo
- convert a 2D floorplan into a 3D model or rendered picture
- build a GLB model of furniture from a product photo
Example prompts for the agent
- Restyle this living room photo into a Scandinavian look with /meltflex:design
- Place this armchair from the product photo into the room with /meltflex:furniture
- Turn this floorplan into a 3D model with /meltflex:3d
How to install MeltFlex AI
Platform CLI
claude mcp add meltflex -e MELTFLEX_API_KEY=<MELTFLEX_API_KEY> -- npx -y meltflex-mcpReplace <…> with your own values
Platform CLI
codex mcp add meltflex-ai --env MELTFLEX_API_KEY=<MELTFLEX_API_KEY> -- npx -y meltflex-mcpReplace <…> with your own values
Config file
{
"mcpServers": {
"meltflex-ai": {
"command": "npx",
"args": [
"-y",
"meltflex-mcp"
],
"env": {
"MELTFLEX_API_KEY": "<MELTFLEX_API_KEY>"
}
}
}
}Replace <…> with your own values
Config file
{
"servers": {
"meltflex-ai": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"meltflex-mcp"
],
"env": {
"MELTFLEX_API_KEY": "<MELTFLEX_API_KEY>"
}
}
}
}Replace <…> with your own values
Platform CLI
gemini mcp add -e MELTFLEX_API_KEY=<MELTFLEX_API_KEY> meltflex-ai npx -y meltflex-mcpReplace <…> with your own values
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about MeltFlex AI
How do I install MeltFlex AI in Claude Code?
Run claude mcp add meltflex -e MELTFLEX_API_KEY=<MELTFLEX_API_KEY> -- npx -y meltflex-mcp, replacing <MELTFLEX_API_KEY> with your own key.
What does MeltFlex AI do?
It's an MCP server for photorealistic interior, facade and garden redesign from a photo, plus furniture-to-3D models, floorplan-to-3D conversion and explorable 3D worlds built from room photos.
What do I need before MeltFlex AI will work?
A personal MeltFlex API key, which only gets issued once there's an active paid subscription on the MeltFlex site. Without that subscription the key won't work.