check_ai_visibilityLABSCO SUMMARY
It answers a question a rank tracker cannot: who does the assistant name?
Type a business, its category and its city, and it comes back with how often that business was named across buying-intent prompts and which competitors appeared in its place. The competitor list is the useful half — it turns “we are invisible” into three names you can go and look at.
What it is
A remote server that tests whether an AI assistant recommends a named local business. You give it a business, a category and a city; it runs buying-intent prompts against one engine and reports a Share-of-Model score alongside the competitors that came up in its place.
What you get
- check_ai_visibility — runs buying-intent prompts for a named business and reports whether it was named, with the competitors that showed up instead
- buying_intent_prompts — generates the prompts a real prospect would ask an AI when looking for that service
- geo_checklist — the GEO/AEO checklist for getting a business named by answer engines
- Detection is heuristic, so the names it reports are a starting point to verify rather than a finished audit
Requirements
A client that accepts a remote Streamable HTTP endpoint, pointed at the vendor's URL. No account and no key.
Setup effort
One command — claude mcp add --transport http local-ai-visibility https://wedge-mcp.andrew7carlin.workers.dev/mcp
USE CASES
01
Finding out whether AI sends your customers elsewhere
Run the check on your own business and read the competitor names that come back. That list is the answer, not the score beside it.
02
Opening a local SEO conversation with a prospect
Check a prospect's business before the call, so the pitch starts from the competitors an assistant already recommends over them.
03
Working out what a customer would actually type
Generate the buying-intent prompts for a category and city, rather than guessing at the phrasing people use when they ask an AI for a recommendation.
04
Deciding what to change next
Pair a poor result with the checklist, so the finding turns into a list of things to fix rather than a number to worry about.
3 tools
Listed in project docs. Your client may show a different list.
geo_checklistbuying_intent_promptsBEFORE YOU INSTALL
What to know first
Tools you get
3 tools named in the project's documentation.
Requires
A client that can reach a remote MCP server over HTTP
Where it runs
All 3 documented clients connect to a hosted endpoint.
INSTALLATION
Connect your client
Server setup
claude mcp add
Claude Code
TERMINAL
claude mcp add --transport http local-ai-visibility https://wedge-mcp.andrew7carlin.workers.dev/mcpHosted endpoint
ENDPOINT URL
https://wedge-mcp.andrew7carlin.workers.dev/mcpAdd this hosted endpoint to your MCP client.
MCP Clients
Cursor
CURSOR · JSON
{
"mcpServers": {
"local-ai-visibility": {
"type": "streamable-http",
"url": "https://wedge-mcp.andrew7carlin.workers.dev/mcp"
}
}
}Any MCP client
ANY MCP CLIENT · JSON
{
"mcpServers": {
"local-ai-visibility": {
"type": "streamable-http",
"url": "https://wedge-mcp.andrew7carlin.workers.dev/mcp"
}
}
}Clients spell the "type" value differently: Claude Code and VS Code use "http", Cline uses "streamableHttp", and Roo Code uses "streamable-http". Claude Desktop does not accept a URL entry in its config file; remote servers go under Settings → Connectors.
