ai-image-merge-mcp
# AI Image Merge MCP
An MCP server for [AI Image Merge](https://aiimagemerge.com). It exposes the product's real image-combination workflow to MCP clients such as Claude Desktop, Cursor, and other compatible agents.
Licensed under the [MIT License](./LICENSE).
Website: https://aiimagemerge.com
## Tools
- `merge_images`: submit exactly two public HTTPS JPG, PNG, or WebP URLs and receive a task id.
- `get_merge_status`: poll a task and receive the generated image URL when it is ready.
The remote service downloads and validates the images, stores them under the authenticated user's storage path, reserves credits, and invokes the configured AI provider. The MCP server never contains an AI provider secret.
## Setup
1. Sign in to [AI Image Merge](https://aiimagemerge.com).
2. Open Settings → API Keys and create a key.
3. Configure the MCP client with:
```json
{
"mcpServers": {
"ai-image-merge": {
"command": "npx",
"args": ["-y", "ai-image-merge-mcp"],
"env": {
"AI_IMAGE_MERGE_API_KEY": "sk-..."
}
}
}
}
```
For a local checkout, replace the command with `node` and the absolute path to `src/index.mjs`.
Optional environment variable:
```text
AI_IMAGE_MERGE_API_URL=https://aiimagemerge.com
```
## Image requirements
The first version accepts two publicly reachable HTTPS image URLs. Each image must be JPG, PNG, or WebP and no larger than 10 MB. Private IPs, localhost, non-HTTPS URLs, and unsupported content types are rejected.
## Development
Run the server directly:
```bash
AI_IMAGE_MERGE_API_KEY=sk-... node src/index.mjs
```
The process speaks newline-delimited JSON-RPC over stdio and keeps stdout reserved for MCP messages.
Run the local protocol smoke test with `npm test`.
## Release and LobeHub listing
This directory is intended to become the root of a separate public GitHub repository, for example `lingn/ai-image-merge-mcp`. After creating that repository and publishing the package to npm, log in to LobeHub and publish the repository with the included manifest:
```bash
npm publish --access public
npx -y @lobehub/market-cli login
npx -y @lobehub/market-cli github connect
npx -y @lobehub/market-cli plugin publish https://github.com/lingn/ai-image-merge-mcp --dir /absolute/path/to/ai-image-merge-mcp
```
`login` and `github connect` open a browser and require the repository owner to complete authorization. The listing's homepage points to `https://aiimagemerge.com`.
TDQS
Scored across 2 tools
merge_images starts a new image merge task, while get_merge_status checks an existing task's status and retrieves the result. Their purposes are clearly distinct and there is no overlap that would confuse an agent.
Both tool names follow a consistent verb_noun pattern: merge_images and get_merge_status. The verbs clearly indicate actions (merge/check) and the nouns describe the target resource.
The server has exactly two operations needed for its narrow purpose: submitting a task and polling for its result. This is well-scoped and does not feel too thin or too heavy.
The tool set covers the full workflow from initiating an image merge to retrieving the generated image URL. There are no obvious missing operations for the stated purpose of merging two images via the AI Image Merge service.