An MCP server that gives your assistant real business data
Six tools an assistant can call directly: find businesses in a city, pull one full record, read any public page as clean text, check whether an email can receive mail, count matches for free, and check its own remaining credits. Hosted by us, so there is no package to install and nothing to keep updated.
POST/mcp
{
"mcpServers": {
"locomint": {
"type": "http",
"url": "https://api.locomint.io/mcp",
"headers": {
"Authorization": "Bearer ${LOCOMINT_API_KEY}"
}
}
}
}
What you get
Every field in the response.
| Field | Meaning |
|---|---|
local_search | Businesses for a category in a city, with a citation on each and text ready to quote |
count_local | How many match, before spending a credit. Free |
place_details | One full record: seven days of hours, contacts, technologies |
read_page | Any public page as markdown or text, with the contacts found on it |
verify_emails | Whether addresses can receive mail, without sending anything |
account_usage | Credits used and left this month. Free |
How it works
What happens between your request and the response.
Hosted, not installed
A URL and your API key. No package on your machine, no version to upgrade, and new tools appear without you doing anything.
Your existing key
Send it as a bearer token. The same credits, quota and rate limit apply as on the REST API, and usage shows in the same dashboard.
Errors the model can act on
A wrong id or a spent quota comes back as a sentence, so the assistant corrects itself instead of failing the conversation.
The same rules
Public pages only, and a site that refuses automated visitors is left alone. Your agent inherits that position.
Pricing
One credit system for every endpoint.
Tools cost exactly what the matching endpoint costs: one credit per business or page, free for counts and usage checks. Nothing extra for using the protocol.
Questions
Things people ask about the MCP Server.
Which assistants can use it?
Any client that supports remote Model Context Protocol servers over HTTP. In Claude Code it is one command; elsewhere it is a short block in a config file, both shown in the documentation.
Do I need a separate key or plan?
No. It is the same key, the same credits and the same dashboard as the REST API. Free plan included.
What happens when my credits run out?
The tool answers with a plain message saying the quota is spent and pointing at the upgrade page, and the assistant relays it rather than inventing data.
