BlipScope
MCP server

The BlipScope index, inside your AI.

Ask Claude, ChatGPT, Cursor or Claude Code which channels are winning in a niche, what a channel is estimated to earn, or who the top earners in a category are — and get the answer from the BlipScope index, labelled the way the app labels it. One URL, OAuth sign-in, no keys to paste.

Three steps, every client

1 — Add

Paste one URL

In Claude: Settings → Connectors → Add custom connector. In Claude Code, one command. Cursor and ChatGPT take the same URL.

https://app.blipscope.com/api/mcp
2 — Approve

Sign in once

Your client registers itself and opens a BlipScope consent screen. Approve it; tokens refresh on their own. Signing out of BlipScope revokes it.

claude mcp add blipscope --transport http https://app.blipscope.com/api/mcp
3 — Ask

Ask in plain English

"Top ten estimated earners in finance." "What does @kurzgesagt earn?" "Channels about woodworking above 50k subscribers." The tools do the lookup; the assistant does the talking.

What it can answer

  • search_channels — channels by keyword, category or minimum subscribers, each with estimated revenue, RPM and outlier score.
  • get_channel — one channel's full record by YouTube channel id.
  • top_channels_by_category — the highest estimated earners in a category.

Estimates travel as estimates. Every field that is modelled says so in its name — estimatedMonthlyRevenueUsd, blipScopeOutlierScore — and every result ends with the provenance line the app prints. An assistant reading the raw output still cannot present a model as YouTube-reported earnings. Subscriber and view counts come from the YouTube Data API; nothing is scraped.

Part of Pro. Connecting is free; tool calls need a BlipScope Pro account — €29 a month billed yearly, unlimited searches, every tool, this server and the Chrome extension included.

Ready?

Put the index where you already work.

Connect the MCP server