@cleanor/mcp
v0.6.1
Published
Cleanor MCP server for AI builders: optimize/convert images, cited storage & image-format data, QR codes, and deterministic dev utilities (hash, uuid, base64, jwt decode, json, color, slugify, count) LLMs get wrong. Run locally over stdio, or use the host
Maintainers
Readme
@cleanor/mcp
Cleanor MCP server for AI builders. The two things most MCP servers don't do: (1) actually optimize and convert the images an AI just generated, and (2) hand back real, cited Cleanor storage and image-format data. Plus QR codes. No API key.
- Tools:
optimize_image,image_format_savings,storage_capacity,qr_code - Hosted (recommended):
https://mcp.cleanor.app/mcp(Streamable HTTP, zero-auth) - Local (this package): runs the same tools over stdio via
npx
Every result links its source on cleanor.app.
Use the hosted server (no install)
If your client supports remote MCP, just point it at the URL:
https://mcp.cleanor.app/mcpRun locally over stdio (this package)
Add it to your MCP client config. It runs on demand via npx:
Claude Desktop / Claude Code
{
"mcpServers": {
"cleanor": {
"command": "npx",
"args": ["-y", "@cleanor/mcp"]
}
}
}Cursor
{
"mcpServers": {
"cleanor": { "command": "npx", "args": ["-y", "@cleanor/mcp"] }
}
}Local image optimization uses sharp (an optional dependency). If sharp is unavailable, the data and QR tools still work; the hosted server uses Cloudflare Images instead.
Links
- Server page: https://cleanor.app/mcp
- Source: https://github.com/cleanor-app/cleanor-mcp
- Web tools: https://cleanor.app/tools
MIT licensed.
