MCP connector
MCP lets Claude read your own organization’s project data directly. Ask in plain language and it picks the tools itself, chains several when the question needs it, and writes the answer in the shape you asked for. The data never leaves your account.
A dashboard answers the questions someone anticipated. The one you need right now usually is not on it.
No tool names or parameters to memorize. Name the project, the period and the platform.
"List the projects we’re tracking." "Which prompts and competitors does this project track?" A project can be named by name, domain or id.
"How has AI visibility trended over the past 30 days, and how does that compare with competitors?" "Which prompts mention us, and which never do?"
"Which keywords dropped the most this week?" "What does the Google top 10 look like for this keyword right now?" Previous rank and competitor positions included.
"Which keywords do we rank well for but AI never mentions us?" A cross-view that neither a pure SEO tool nor a pure GEO tool can give you.
"Research related keywords for this term and pull out the ones with buying intent." "Build an outline for this keyword, then write it up as a 1,500-word article."
This is the real difference from opening a dashboard: you describe the goal, not the steps.
The MCP chapter of the user manual lists 42 ready-to-use example questions, grouped by inventory, visibility, competitors, gaps and content production.



The question worth asking about any AI connector is what it can read — and what it cannot.
Every tool is scoped to the organization the key belongs to. Other organizations’ projects, rankings and usage are not visible.
Each member holds their own. Regenerating invalidates the old one immediately. Both the member and the admin can delete it.
Changing a password — self-service, admin reset or forgot-password — automatically revokes that member’s key. A new one has to be created.
MCP can be switched off for the whole organization at any time. Revoking one member’s access revokes their key at the same time.
QWhat is MCP?
Model Context Protocol — an open protocol for connecting AI assistants to external data sources. Once connected, Claude reads the system data you have authorized instead of relying on you to paste it in.
QDo we need to write any code?
No. On claude.ai, go to Settings → Connectors, paste the URL, sign in and confirm your password. Claude Code and Claude Desktop connect with a key — also a single command.
QCan Claude see our other clients’ data?
No. The key is bound to an organization and every query is scoped to that organization’s projects. Isolation is enforced at the system level, not by instructions in a prompt.
QDoes it consume plan quota?
Yes. Read-only queries cost very little; research, live fetching and content generation share the same quota as the web tools, priced the same way. Once connected you can simply ask how much you have used this month.
QHow do we get started?
An admin enables it on the MCP Keys page, and members then create their own key — or connect with one-click authorization on claude.ai. To see it working first, we can connect it to your projects during a demo.
Our team will get back to you as soon as possible to schedule a live demo.