# pedra-ai/pedra-mcp [Health: Active]

**Category:** 🏠 Real Estate  
**Repository:** https://github.com/pedra-ai/pedra-mcp  
**GitHub Stars:** 1  
**npm Downloads (last month):** 210  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/pedra-ai-pedra-mcp

## Description
AI photo and video editing for real-estate listings via the Pedra API: virtual staging, renovation, room emptying, photo enhancement, sky replacement, object removal/blur, and property video generation. npx @pedra-ai/mcp

## Tools
Capabilities this server exposes over MCP:

- **pedra_enhance** — Improve lighting, color, sharpness
- **pedra_enhance_and_correct_perspective** — Enhance + straighten perspective
- **pedra_empty_room** — Remove all furniture/objects
- **pedra_furnish** — Virtually stage a room
- **pedra_renovation** — Renovate walls/floors/finishes
- **pedra_edit_via_prompt** — Edit from a natural-language prompt
- **pedra_sky_blue** — Replace a dull sky with clear blue
- **pedra_remove_object** — Remove an object using a mask
- **pedra_blur** — Blur faces, license plates, etc.
- **pedra_create_video** — Render a property video from images
- **pedra_update_video** — Edit a video without re-rendering unchanged clips
- **pedra_generate_voice_script** — Write a voiceover script from property photos
- **pedra_generate_voice** — Turn a script into a voiceover audio track
- **pedra_music_library** — List background-music tracks + voice languages
- **pedra_list_properties** — List the account's properties
- **pedra_list_property_images** — List a property's photos as URLs
- **pedra_create_property** — Create a property
- **pedra_add_images_to_property** — Add photos to a property by URL
- **pedra_credits** — Read plan + remaining credits
- **pedra_feedback** — Thumbs up/down + optional credit-back

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "pedra-mcp": {
    "command": "npx",
    "args": ["-y","@pedra-ai/mcp"],
    "env": {
      "PEDRA_API_KEY": ""
    }
  }
}
```

**Requires environment variables:** `PEDRA_API_KEY` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What pedra-ai/pedra-mcp MCP server does

The pedra-ai/pedra-mcp MCP server connects MCP clients to Pedra's AI real-estate media API. It provides separate tools for common listing workflows: improving lighting and sharpness, correcting perspective, emptying rooms, virtual staging, renovation, prompt-based editing, sky replacement, object removal, and privacy blurring. It also supports property video creation and updates, voiceover script generation, voice synthesis, music and voice-language lookup, and basic property management.

Property management tools can list the account's properties, retrieve a property's image URLs, create a property, and add images by URL. Account tools report the current plan and remaining credits, while the feedback tool accepts a thumbs-up or thumbs-down and can optionally request credit back.

## How it works

The pedra-ai/pedra-mcp MCP server is a thin wrapper around the Pedra SDK and maps one MCP tool to each API endpoint. Calls are blocking: image operations return their resulting URL or URLs directly, and video creation waits for the completed video rather than exposing a job ID for client-side polling. Video generation can take up to approximately 10 minutes, while the API connection remains active with a heartbeat.

Most image operations accept an `imageUrl` and optional parameters. Image, mask, and video-frame inputs may be public HTTPS URLs, `data:` URIs, or absolute paths to local files. For local files, the server reads and inlines the image as base64. Supported formats include JPG, JPEG, PNG, WebP, GIF, BMP, TIFF, HEIC, and AVIF, with a stated maximum size of 40 MB. An image pasted directly into chat is not automatically treated as a file path; save it or provide a file URL instead.

API errors such as insufficient credits or invalid images are returned as readable MCP tool errors. The server itself does not store data. URLs and parameters supplied to the tools are sent to the Pedra API, whose privacy policy covers data handling, retention, and sharing.

## Setup and configuration

The pedra-ai/pedra-mcp MCP server requires a Pedra API key obtained from a Pedra account. Set that credential in the `PEDRA_API_KEY` environment variable. The published npm package can run without a global install:

```bash
PEDRA_API_KEY=your-api-key npx -y @pedra-ai/mcp
```

For clients that use JSON configuration, set the command to `npx`, use `-y @pedra-ai/mcp` as the arguments, and provide `PEDRA_API_KEY` in the server environment. The documented configuration works with Claude Desktop and Cursor. A bundled `.mcpb` release is also available for Claude Desktop, where installation prompts for the API key. The server can be configured through other MCP clients as well.

The API key is required for requests to Pedra. Tool-specific input fields and optional parameters are exposed through each MCP tool's schema, so clients can inspect those schemas when constructing calls.

_Full upstream README: https://allmcps.com/mcp/pedra-ai-pedra-mcp/readme_