Skip to content
Fedlum is in private beta for brands and agencies in Mexico.Talk to us
fedlum
Sign up
MCP server

Ask your AI assistantabout your AI visibility.

Fedlum's MCP server gives Claude, ChatGPT, Cursor or VS Code access to your projects, reports and answers, with the same numbers you see in the dashboard. Ask in Spanish or English.

Create your tokenRead the guide
Available on every paid plan.
What it is

A bridge between your data and your agents

MCP (Model Context Protocol) is an open standard that lets AI assistants use external tools. With Fedlum's server, your assistant can look up your metrics, read full answers and add prompts without you copying and pasting anything.

  • The same queries that power the dashboard, the API and exports
  • Respects your permissions: you only see the projects you can access
  • Write tools only work for owners
Tools

Ten tools, from lookup to action

Your assistant picks the right one based on what you ask. These are the ones the server exposes today.

Your projectRead-only

  • list_projectsThe projects you have access to.
  • list_brandsThe brands a project tracks, starting with yours.
  • list_topicsA project's topics.
  • list_model_channelsThe AI models the project measures.
  • list_promptsActive, suggested or archived prompts.

ReportsRead-only

  • get_brand_reportVisibility, Share of Voice, sentiment, position and win rate per brand, broken down by date, model, prompt, topic or country.
  • get_domain_reportThe domains models used as sources, with retrieval and citation metrics.

AnswersRead-only

  • list_chatsCollected answers and the brands they mention.
  • get_chatOne full answer: the prompt, the text, brands, sources and fan-outs.

WriteWrites data

  • create_promptsAdds up to 50 prompts to a project in one go.
Ready-made workflows

Seven workflows to get started

Instruction templates your assistant can run in one click. They combine several tools into a complete analysis.

  • weekly_pulse

    Weekly pulse

    The last 7 days against the previous 7: what changed, on which model, and three actions.

  • competitor_radar

    Competitor radar

    Where a competitor gained ground in the last 30 days, by prompt and by model.

  • engine_scorecard

    Engine scorecard

    Your brand on each AI model, strongest to weakest, with what explains the gaps.

  • topic_heatmap

    Topic heatmap

    Visibility by topic for every brand, and the topics where you trail.

  • prompt_grader

    Prompt grader

    An A–F grade for each prompt and what the winning brand does differently.

  • source_authority

    Source authority

    The 15 domains that carry the most weight in your category, grouped by type, and which to target.

  • campaign_tracker

    Campaign tracker

    Your visibility 14 days before and after a date, to tell whether a campaign moved anything.

Setup

Connect it in three steps

  1. 1

    Create a personal token

    In Fedlum, go to Settings › API keys and create one under Personal tokens (MCP). It starts with fl_pat_.

  2. 2

    Add the server to your client

    In Claude, ChatGPT, Cursor or VS Code, add a remote MCP server with this address and your token in the authorization header.

    https://api.fedlum.com/mcp
  3. 3

    Ask

    Try “Which sources cite my competitor but not us this month?” or run one of the ready-made workflows.

  4. Treat your token like a password: don't share it or commit it to a repository. You can revoke it anytime.
  5. Sign-in with your Fedlum account (OAuth), no token neededComing soon
Example configuration
{
  "mcpServers": {
    "fedlum": {
      "url": "https://api.fedlum.com/mcp",
      "headers": {
        "Authorization": "Bearer fl_pat_…"
      }
    }
  }
}

See where your brand stands today

Create your project, pick your prompts and get your first data in minutes. If you'd rather have us walk you through it, book a call.

Start free trialTalk to sales

7-day free trial · No card required

FAQ

Questions we hear a lot

Which AI clients work?

Any client that supports remote MCP servers over HTTP, such as Claude, ChatGPT, Cursor or VS Code.

Which plan do I need?

The MCP server is included in every paid plan, for brands and agencies.

Do the numbers match the dashboard?

Yes. The server uses the same queries as the dashboard, the API and exports, so a report returns the same result everywhere.

What can my assistant do with my data?

Only what your role allows. Read tools respect your access, and the only tool that writes, adding prompts, requires being a project owner.

Can I ask in Spanish?

Yes. Tools are described in English so agents understand them, but you can ask and get answers in Spanish.