Suparanku MCP: bring recommendations into a compatible AI client

What MCP is, in one sentence

MCP (Model Context Protocol) is an open standard, introduced by Anthropic in late 2024, that lets a compatible AI client connect to an outside service and use its data and tools. Client support is not universal: availability, plan requirements and setup differ by client.

Suparanku exposes one remote MCP endpoint. Subject to authorization, a compatible client can retrieve visibility, position, sentiment, competitor and cited-source data, plus the recommendations produced by the Growth engine.

Why this is genuinely useful, not just a buzzword

MCP can bring Suparanku data into a compatible client the team already uses. A request such as “what changed since the previous AI-visibility scan?” can return the available metrics in that client without a CSV export.

Repeatability matters because brand recommendations vary between runs. In 2,961 runs across ChatGPT, Claude and Google’s AI, SparkToro found less than a 1-in-100 chance of the same brand list appearing twice and roughly a 1-in-1,000 chance of the same order. A single list is not a fixed ranking; repeated samples and a stable prompt set provide the useful context.

A simple example

Consider Tanaka-san, a marketing lead whose president has started asking, “are we showing up in ChatGPT?” At the June 2026 launch, Starter MCP access was read-only. Full Starter read/write access arrived with the July repricing; today every paid plan defaults to full MCP access. A weekly digest still needs only a read-only credential.

1. Connect a compatible client. In Settings → MCP & API tokens, create a Personal Access Token for a client that accepts one, or use OAuth where the client supports it. The external client may require its own paid plan or administrator approval.
2. Every Monday morning, one message: “From Suparanku, pull this week’s visibility changes for our brand and the top 3 recommendations, then summarize it in three lines for my manager.”
3. Review the result. The client calls Suparanku’s read tools, retrieves the latest available metrics and automatically generated recommendations, and drafts a summary for review.

The recommendations come from Suparanku; MCP delivers them to another working surface. If the external client supports scheduled tasks, and the user’s client plan and administrator settings allow them, the same prompt can be scheduled there. Scheduling is a client feature, not a Suparanku MCP guarantee.

The practical value is simple: retrieve the analysis already available in Suparanku without exporting a spreadsheet.

What you can pull — and what needs write access

With a read-only credential, a compatible client can, for example:

• Read the 5-axis analytics: visibility, position, sentiment, competitors and cited sources.
• Read available recommendations and their evidence.
• Browse cited sources and available scan history.

Current paid plans default to read/write MCP access, but that does not make every call available to every user. Write tools require company access set to full, an eligible subscription state, a suitable membership role and a read/write credential. Quotas and plan gates still apply. Destructive tools additionally require an explicit destructive-action scope and per-call confirmation.

Connecting is safe by design

MCP authorization combines several controls:

• Company, role and subscription checks. Tenant isolation and membership are re-evaluated on every call. Read-only credentials cannot write, and plan or subscription gates can narrow the available operations.
• Credential scopes. Personal Access Tokens can be limited to selected companies and read-only access. OAuth grants can be downscoped to read access; their availability depends on the client.
• Explicit destructive controls. Delete operations need a destructive-action scope and confirm: true, in addition to the normal role and plan checks.
• Revocation. Owners and administrators can block a credential’s access to their company from Settings → History & Access.

Try it

If you use a compatible client, open Settings → MCP & API tokens and choose the authorization method that client supports. The always-current tool reference is at app.suparanku.com/docs/mcp; the versioned endpoint is app.suparanku.com/api/mcp/v1.