Model Context Protocol (MCP) Server
Turn Useful Utilities into a native toolkit for your AI agents. Connect Claude Desktop, Cursor, or your custom AI to our MCP endpoint to sanitize filenames, detect homoglyphs, format JSON, generate passwords, convert currencies, validate GST numbers, and more โ 63 tools in total.
What is the MCP Server?
The Model Context Protocol (MCP) is an open standard that enables AI models to securely connect to external data sources and tools. Instead of manually pasting data into our web interface, your AI agent can call our tools autonomously behind the scenes.
Available Tools
63 tools across 10 categories โ your MCP client discovers the full list, with exact names and parameters, automatically once connected. A few highlights:
sanitize_filename: Automatically removes forbidden characters and handles Windows reserved names to preventInvalidCharError.detect_homoglyphs: Scans strings for Unicode lookalike characters, zero-width spaces, and invisible markers used in spoofing attacks.format_json: Validates, prettifies, or minifies JSON strings. Can recursively sort JSON keys.time_calculate_difference: Precisely calculates the duration between two ISO-8601 date strings down to the millisecond.security_generate_password: Generates passwords, memorable passphrases, or PINs with an entropy/strength rating.wealth_convert_currency: Converts between currencies using live exchange rates, with an offline fallback.- Plus list, table, text, JSON, SQL/CSS, and loan/debt utilities โ the full catalogue mirrors the site's text-based tools.
How to Connect (Claude Desktop / Cursor)
We provide an SSE (Server-Sent Events) endpoint. You can connect to it using the official MCP Inspector proxy, which bridges stdio to our remote web server.
Add the following configuration to your claude_desktop_config.json or Cursor settings:
{
"mcpServers": {
"useful-utilities": {
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/inspector",
"https://mcp.usefulutils.in/sse"
],
"env": {
"AUTHORIZATION": "Bearer YOUR_FREE_API_KEY"
}
}
}
}Connect to a focused subset (recommended)
Your client loads every tool definition into the model's context and re-sends it on every request for the whole session. The full catalogue costs roughly 12,000 tokens โ about 6% of a 200K context window โ which is a poor trade if your agent only ever reaches for a handful of tools.
So each category is also served on its own endpoint. Point your client at the one you actually want and pay a fraction of the context:
| Endpoint | Tools | Context cost | Contains |
|---|---|---|---|
/mcp/dev | 10 | ~1.2% | package versions, regex, cron, SQL, CSS, filenames, JSON |
/mcp/files | 9 | ~0.7% | PDF merge/split, image resize/compress/convert/crop/redact, OCR |
/mcp/data | 33 | ~2.7% | JSON, lists, tables, text, time |
/mcp/finance | 8 | ~0.9% | currency, loans, debt, ROI, GSTIN, PAN, IFSC |
/mcp/security | 3 | ~0.5% | hashing, password generation, email headers |
/mcp | 63 | ~6.0% | everything |
Same API key everywhere, and /mcp and /sse are unchanged โ an existing configuration keeps working exactly as before.
Rate Limits & API Keys
To protect our serverless infrastructure, access to the MCP server requires a free API key.
- Free Tier Limit: 200 tool calls per day per API key.
- Data Privacy: The content of your tool calls โ arguments and results โ is processed in-memory ephemerally and is never stored or logged.
- What we do record: when you create a key, we store the key, the IP address it was created from, and an approximate country, state and city derived from it, purely to enforce fair use and prevent abuse. Usage analytics record which tool was called, never what you passed to it. See the privacy policy for the full list.
Free, instant, no signup โ limited to 3 new keys per IP per day.