MCP ServerStreamable HTTP28 tools

Geolix MCP Server

Connect Claude, Cursor and other MCP clients straight to your Geolix data. Ask in plain language; the answers come from the same data the dashboard shows.

Claude Code
claude mcp add geolix --transport http https://api.geolix.ai/mcp \
  --header "x-api-key: glx_live_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
server url
https://api.geolix.ai/mcp
transport
Streamable HTTP
auth
x-api-key header, the same key as the REST API
tools
2818 read-only, 10 that change configuration; clients ask before running a write tool.
clients
Claude Code · Claude Desktop · Cursor · VS Code · Windsurf

Overview

MCP (Model Context Protocol) is the open protocol AI assistants use to call external tools. Once connected, your assistant can look up brand visibility, compare competitors, inspect citation sources and add monitored intents on your behalf. You just ask.

Setup

01

Create an API key

Sign in at app.geolix.ai, open Settings → API keys, create a key and copy it. MCP and the REST API share the same kind of key.

02

Add the server to your client

Every client needs the same three things:

SettingValue
Server URLhttps://api.geolix.ai/mcp
TransportStreamable HTTP
Headerx-api-key: glx_live_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

Where to put them in each client is in the next section.

03

Ask

Once connected, just ask, for example: "How did our visibility in ChatGPT change over the last 30 days?" The assistant picks the tools, pulls the data and answers.

Client configuration

Claude Code

claude mcp add geolix --transport http https://api.geolix.ai/mcp \
  --header "x-api-key: glx_live_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"

Claude Desktop

Settings → Connectors → Add custom connector, paste the server URL; for authentication choose header and enter x-api-key.

Cursor

Settings → Tools & Integrations → MCP → Add Custom MCP, or edit .cursor/mcp.json:

{
  "mcpServers": {
    "geolix": {
      "url": "https://api.geolix.ai/mcp",
      "headers": { "x-api-key": "glx_live_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx" }
    }
  }
}

VS Code (GitHub Copilot)

Create .vscode/mcp.json; keep the key in an environment variable rather than committing it:

{
  "servers": {
    "geolix": {
      "type": "http",
      "url": "https://api.geolix.ai/mcp",
      "headers": { "x-api-key": "${env:GEOLIX_API_KEY}" }
    }
  }
}

Windsurf

Settings → MCP → Add Server, enter the server URL and the x-api-key header.

What to ask

A few prompts that work as-is. The assistant translates them into calls from the tool list below.

How it works

On connect the assistant fetches the tool list (name, parameters, whether it is read-only). For each question it picks tools, calls them with arguments, and the server queries the Geolix database and hands JSON back for the assistant to compose an answer.

The data is exactly what the dashboard shows: the MCP server and the REST API read the same database under the same permissions, with no extra model processing in between.

Permissions

18 of the 28 tools are read-only and only pull data; 10 change configuration (create brands, add intents, trigger or cancel runs and so on).

Write tools are annotated with readOnlyHint: false, archive-type tools also carry destructiveHint: true; clients that honour these hints ask for confirmation before running them.

noteWrite tools require the key owner to be an organisation owner or admin; read tools work for any member.

Tools

Grouped by resource. The tool name is what the assistant calls.

Brands

ToolAccessDescription
list_brandsreadList brands
get_brandreadBrand details
create_brandwriteCreate a brand
update_brandwriteUpdate a brand

Projects & intents

ToolAccessDescription
list_projectsreadList projects
get_projectreadProject details
create_projectwriteCreate a project
update_projectwriteUpdate project settings
list_queriesreadList monitored intents
add_querieswriteAdd monitored intents
update_querywriteUpdate an intent
archive_querywriteArchive an intent
suggest_topicsreadAI-suggested topics to monitor
suggest_queriesreadAI-suggested intents

Runs

ToolAccessDescription
list_runsreadList detection runs
get_runreadRun details and progress
get_completionreadCompletion of the current window
trigger_runwriteTrigger a run
cancel_runwriteCancel a run

Board

ToolAccessDescription
get_boardreadOverview metrics
get_board_sourcesreadCitation sources
get_source_detailreadDetail for one source
get_competitive_gapreadCompetitive gap
get_by_intentreadMetrics by intent
get_chat_recordsreadRaw question-and-answer records
get_chat_facetsreadFilter facets for records
export_chatsreadExport records

Others

ToolAccessDescription
reanalyze_projectwriteRe-aggregate a project's history

Troubleshooting

401 unauthorized

403 forbidden

Connection timeout