Mencoro MCP server

Ask an AI assistant how your brand is doing in AI answers — rank, mentions, sentiment and Share of Voice — and have it set up and tune your monitoring, in plain language.

Mencoro tracks how brands surface in AI answer engines — ChatGPT, Perplexity, Google AI Overview and AI Mode — and in Google Search and Shopping. The MCP server exposes that data to any MCP client, and lets it manage projects, tracked queries, clusters and organizations, as 60 tools (31 that read, 29 that change something) and 20 prompts. One of them, get_mencoro_guide, answers questions about Mencoro itself and works without signing in.

The server is hosted. Most clients should connect to it directly:

https://api.mencoro.com/mcp

This repository holds the public metadata for that server plus a small stdio bridge (@mencoro/mcp) for hosts that can only launch a local process. It does not contain the Mencoro application source.

Connect

Clients that speak remote MCP (recommended)

No install, no local process. Sign in with OAuth, or paste a personal access token as a header.

Clients that can only spawn a local process

Claude Desktop's manual configuration, and older stdio-only hosts, need a bridge. That is what this package is:

npx -y @mencoro/mcp

// claude_desktop_config.json
{
  "mcpServers": {
    "mencoro": {
      "command": "npx",
      "args": ["-y", "@mencoro/mcp"],
      "env": { "MENCORO_API_KEY": "mcp_pat_your_token" }
    }
  }
}

The token goes through the environment rather than an argument, so it does not appear in ps output and does not have to survive the host's argument splitting.

Authentication

Two ways in. Both carry the same three permissions, and neither can ever do more than your own role in an organization allows:

OAuth 2.1 — the one-click path. Supported by Claude, ChatGPT, Claude Code and any client that implements the MCP authorization spec. Nothing to copy or paste; revoke it from the app. The consent screen pre-selects the scopes the client asked for (all three when it asked for none), and you can untick any but read. The server advertises PKCE (S256), Client ID Metadata Documents, Dynamic Client Registration and RFC 9728 resource metadata, so clients discover everything they need from https://api.mencoro.com/.well-known/oauth-protected-resource/mcp. Calling a tool the connection lacks the scope for answers 403 insufficient_scope naming the scopes to request, so a client that supports step-up authorization asks you to reconnect with the extra permission.

Personal access token — for CLI clients, config files, and this bridge. Create one at tool.mencoro.com/me/mcp-server; it is shown once, starts with mcp_pat_, carries the permissions you tick (all three by default), and can be pinned to a single organization and given an expiry. Send it as Authorization: Bearer mcp_pat_…. A call the token lacks the permission for fails with a tool error naming it; a token cannot be upgraded, so create a new one to add a permission.

Connections and tokens created before write access existed have read and write but not organization:manage.

ChatGPT cannot send a custom Authorization header to a remote connector — use OAuth there.

Client configuration

{
  "mcpServers": {
    "mencoro": {
      "url": "https://api.mencoro.com/mcp",
      "headers": { "Authorization": "Bearer mcp_pat_your_token" }
    }
  }
}

{
  "servers": {
    "mencoro": {
      "type": "http",
      "url": "https://api.mencoro.com/mcp",
      "headers": { "Authorization": "Bearer mcp_pat_your_token" }
    }
  }
}

[mcp_servers.mencoro]
url = "https://api.mencoro.com/mcp"
http_headers = { "Authorization" = "Bearer mcp_pat_your_token" }

{
  "mcpServers": {
    "mencoro": {
      "httpUrl": "https://api.mencoro.com/mcp",
      "headers": { "Authorization": "Bearer mcp_pat_your_token" }
    }
  }
}

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "mencoro": {
      "type": "remote",
      "url": "https://api.mencoro.com/mcp",
      "enabled": true,
      "headers": { "Authorization": "Bearer mcp_pat_your_token" },
      "oauth": false
    }
  }
}

{
  "mcpServers": {
    "mencoro": {
      "serverUrl": "https://api.mencoro.com/mcp",
      "headers": { "Authorization": "Bearer mcp_pat_your_token" }
    }
  }
}

{
  "mcpServers": {
    "mencoro": {
      "type": "http",
      "url": "https://api.mencoro.com/mcp",
      "headers": { "Authorization": "Bearer ${MENCORO_API_KEY}" }
    }
  }
}

Tools

31 tools read and 29 change something. Every tool declares all four MCP annotations (readOnlyHint, destructiveHint, idempotentHint, openWorldHint) explicitly, so a client can decide what to ask you before calling it.

Reading

Every read tool needs only the read scope.

Changing things

A tool marked C needs a confirmation: call preview_operation with the tool's name and arguments, show the plan to the user, and call the tool with the returned confirmationToken only after they agree. The token is single-use, expires after a few minutes, and is refused if anything the plan described has changed. Creating tools, job starters and confirmable tools also accept a requestId: a retry with the same one returns the first result instead of doing the work twice.

A change is refused exactly where the Mencoro app refuses it: your role in the organization, an archived project, or a missing subscription where the app requires one.

Dates are ISO YYYY-MM-DD and must fall inside the retention window. Positions are 1-based and lower is better; every other metric improves as it rises.

The live definitions — names, descriptions, input and output schemas, annotations — are readable without a credential at https://api.mencoro.com/public/v1/mcp, a discovery-only mount of the same server that answers initialize, ping, tools/list, prompts/list and the public guide (get_mencoro_guide), and refuses everything else. Calling any other tool requires signing in at https://api.mencoro.com/mcp.

Prompts

Twenty ready-made starting points, surfaced by clients that support MCP prompts. Twelve ask about your data:

brand_ai_overview · whats_changed · organization_overview · top_queries · biggest_movers · query_history · competitor_standing · head_to_head · negative_mentions · cited_sources · coverage_health · sov_explainer

Four script a change step by step, stopping for your approval where it matters:

Four turn the data into decisions:

ChatGPT submission and skills

chatgpt-app-submission.json contains the app submission metadata, tool annotation justifications, and review test cases. Keep its tool descriptions and justifications aligned with the hosted server when capabilities change.

Ten optional skills turn the tools into reusable workflows. ChatGPT does not surface MCP prompts, so these are how its users get the same guided flows.

Six analyse, and change nothing:

• Visibility report: performance summaries, trends, and query gains or losses.
• Competitor analysis: share of voice and head-to-head comparisons.
• Sentiment review: sentiment breakdowns with attributed mention excerpts.
• Coverage audit: current monitoring coverage and stale queries.
• Results review: whether a change moved AI visibility, before and after against a baseline.
• Optimization insights: a prioritised list of what to improve, optionally with connected Ahrefs, Semrush, Search Console or Google Analytics tools.

Four make changes, each only after the user approves it, and need the write scope:

• Project setup: a monitored project from a website in a short conversation, with brand names and suggested competitors proposed before anything is created.
• Query expansion: new prompts or keywords worth tracking, added as the user picks them.
• Cluster reorganisation: a proposed clustering, reviewed and applied.
• Cost tuning: frequency, pass and pause changes that bring check spend within the plan.

In the submission portal's Skills step, upload each skill folder under skills/, or a ZIP containing that folder. Include both SKILL.md and agents/openai.yaml; the latter declares the existing Mencoro MCP connection. Users need to connect their Mencoro account through OAuth. These skills require no additional backend endpoint or local executable.

The portal stores an uploaded snapshot. Upload revised bundles when instructions change, and test each workflow with a connected account before submitting the app. See the OpenAI skills guide.

The bridge

Run it

export MENCORO_API_KEY=mcp_pat_your_token
npx -y @mencoro/mcp            # serve on stdio
npx -y @mencoro/mcp doctor     # check the endpoint, the token, and list the tools

Docker

docker run --rm -i -e MENCORO_API_KEY=mcp_pat_your_token ghcr.io/mencoro/mencoro-mcp

-i is required and -t must be omitted: the MCP transport is this process's stdin and stdout.

Options

What it actually does

It splices your client's stdio transport onto a Streamable HTTP transport and forwards every JSON-RPC frame verbatim, in both directions. The only frame it looks inside is the handshake — enough to echo the negotiated protocol version back upstream and to rebuild the session if the server evicts it, which a deploy or a scale-down will do. It knows no other method, so it cannot drift from the server: tools, prompts, resources, completions, progress notifications and anything added later all pass straight through.

Without a credential

The bridge starts anyway, serving catalog.json — a committed copy of what the hosted server advertises — plus a mencoro_setup tool. So npx -y @mencoro/mcp introspects to the same tools and prompts a credentialed run does, and anything that scans the package sees a real catalogue rather than a server that looks empty. Calling one of those tools returns the setup instructions as an error; it never returns invented data. The exception is get_mencoro_guide, which reads only published material: the bridge forwards it to the hosted server's anonymous endpoint and returns the real answer.

npm run sync:catalog refreshes the file from the anonymous catalogue endpoint, and npm run check:catalog fails if the committed copy has fallen behind. A scheduled workflow runs the check weekly, because the catalogue it mirrors lives in another repository and nothing in a pull request here would notice it drifting.

Limits

Repository contents

Development

npm ci
npm run typecheck
npm test            # builds first, then runs the suite
npm run check:manifests
npm run check:catalog   # asks the hosted server whether catalog.json is still current

Tests run the TypeScript sources directly through Node's type stripping, so development needs Node 22.18 or newer. The published package targets Node 20.19+, which CI verifies separately against the built artifact.

Releases are cut by tagging. npm version <patch|minor|major>, mirror the new version into server.json (.version, the npm package entry, and the image tag), run npm run check:manifests, then push the tag — CI publishes to npm, GHCR and the MCP registry, in that order, then creates the GitHub release, which is what Glama builds a new version from.

Support

• Product and account questions — support@mencoro.com
• Bugs in this package — open an issue
• Security — see SECURITY.md
• About the server — mencoro.com/features/mcp-server
• Privacy policy — mencoro.com/legal/#privacy
• Terms of service — mencoro.com/legal/#terms

Licence

MIT. See LICENSE.

The Mencoro name, logo and brand assets in assets/ are trademarks of Mencoro and are not covered by that licence.