# strato-space/media-gen-mcp [Health: Active]

**Category:** 🎥 Multimedia Process  
**Repository:** https://github.com/strato-space/media-gen-mcp  
**GitHub Stars:** 9  
**npm Downloads (last month):** 368177  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/strato-space-media-gen-mcp

## Description
TypeScript MCP server for OpenAI Images/Videos and Google GenAI (Veo) media generation, editing, and asset downloads.

## Tools
Capabilities this server exposes over MCP:

- **fetch-images**
- **test-images**
- **resource_link**
- **ResourceLink**
- **image**
- **ImageContent**
- **resource**
- **EmbeddedResource**
- **resource.blob**
- **fetch-document**
- **structuredContent**
- **MEDIA_GEN_MCP_URL_PREFIXES**

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

```json
"mcpServers": {
  "media-gen-mcp": {
    "command": "npx",
    "args": ["-y","github"]
  }
}
```

## Documentation

## What strato-space/media-gen-mcp MCP server does

strato-space/media-gen-mcp MCP server provides MCP tools for creating and editing images, generating videos, managing asynchronous video jobs, and retrieving media assets. OpenAI image operations support the `gpt-image-1.5` and `gpt-image-1` models. The image workflow covers text-to-image generation and edits such as inpainting, outpainting, and compositing, with support for one to 16 input images.

For video, the server supports OpenAI Videos with Sora models, including job creation, remixing, listing, retrieval, deletion, and content downloads. Google GenAI support uses Veo video generation operations, including starting an operation, polling its status, and downloading completed MP4 output. The server also includes tools for fetching images, videos, and documents from HTTP(S) URLs or local filesystem paths.

## How it works

The server is implemented in strict TypeScript and returns MCP `CallToolResult` objects. Depending on the tool and response options, media can be returned as inline `image` or `resource` content, as `resource_link` entries pointing to files on disk, or through structured output. Image responses can be placed in `content`, `structuredContent`, or at the top level for testing and client compatibility checks.

Downloaded and generated assets are written to disk with generated filenames. Image and document names include a generated identifier, while video names can use the OpenAI video ID. For images, the server can automatically choose file-based output when an inline response would exceed a safe size threshold. Optional Sharp-based processing can compress images by reducing quality and dimensions.

## Setup and configuration

The repository provides a local installation flow based on cloning the project, installing its npm dependencies, and building the TypeScript source:

```sh
git clone https://github.com/strato-space/media-gen-mcp.git
cd media-gen-mcp
npm install
npm run build
```

The standard build performs strict type checking, including library checks. An `npm run esbuild` mode is also documented for faster bundling without type checking. The supplied material does not specify the credential variable names, client configuration JSON, or the command used to launch the built server, so those details should be checked in the repository before deployment.

## Tools and capabilities

The strato-space/media-gen-mcp MCP server includes these documented capability groups:

- Generate and edit OpenAI images with model, size, quality, background, moderation, format, compression, count, and user controls.
- Create, remix, list, retrieve, delete, and download OpenAI Sora video jobs.
- Start, poll, and download Google Veo video operations.
- Fetch and optionally compress up to 20 images in parallel from URLs or local paths.
- Fetch videos and documents, returning file links or embedded resources.
- Test MCP image result placement with the `test-images` tool.
- Report validation, API, and filesystem failures with `isError: true` and a text error item.

## Limitations and notes

The server depends on the OpenAI and Google GenAI APIs for the corresponding generation features, so model availability and provider-side behavior apply. Video generation may be asynchronous; the OpenAI create tool and Google generation tool can optionally wait, while separate retrieval tools handle existing jobs or operations.

Inline media responses may not be suitable for every MCP client or payload size. The server therefore offers file-backed `resource_link` and embedded `resource` forms, but clients must support the selected MCP result shape to consume them correctly. Image compression is user-controlled and disabled by default. The provided material also indicates that DALL·E support is planned rather than describing it as an available generation path.

_Full upstream README: https://allmcps.com/mcp/strato-space-media-gen-mcp/readme_