Directory coverage

A beautiful MCP endpoint with zero directory presence is a dead launch

Building the server is roughly forty percent of the job. Directory coverage is the coordinated sprint across consumer stores, developer IDEs, the official MCP Registry, MCPBundles /skills, and community indexes — so humans and agents actually find mcp.yourbrand.com.

Where agents actually look

Coverage is a matrix, not a single listing. These four facts decide how much of it you hold after launch.

Three discovery modes

Consumer shelf (ChatGPT, Claude), developer IDE (Cursor, Registry, VS Code), and open-web agent search (PulseMCP, Smithery, Glama, MCPBundles /skills, docs crawl). Each mode uses different directories — one launch sprint covers all rows.

Community indexes + MCPBundles

PulseMCP, Smithery, Glama, and MCPBundles /skills are the same sprint row — public skill page, Hub connect path, IndexNow, and cross-list refresh alongside third-party community directories.

Registry propagates

Official MCP Registry publish on customer DNS feeds VS Code, Copilot curation, and crawlers that mirror the canonical index — put geo keywords in the server name, not only the description.

24 discovery checkpoints

Launch sprint covers consumer stores, Registry, Cursor, MCPBundles /skills, and community indexes — not just mcp.yourbrand.com going live. Most brands skip IDE and registry rows until someone coordinates the full matrix.

Directory tiers we run by default

Exact checklist varies by package — Launch + Widget and Growth include the full consumer + developer sweep.

  • Tier A — ChatGPT Apps and Claude Connectors (consumer shelf)
  • Tier B — Cursor marketplace plugin, cursor.directory, IDE quick-start docs
  • Tier C — Official MCP Registry DNS publish on customer domain
  • Tier D — PulseMCP, Smithery, Glama, MCPBundles /skills, Hub listing, and CLI discoverability
  • Tier E — Customer /docs, llms.txt, footer PR snippets (required for directory success)

How the sprint runs

  1. 1

    Kickoff + checklist

    Confirm directory goals, competitor shelf presence, and which tiers belong in scope — stores, Registry, Cursor, /skills, community dirs, docs.

  2. 2

    Submit in dependency order

    Registry and remote MCP health first; consumer store JSON; Cursor plugin pointing at mcp.yourbrand.com/mcp.

  3. 3

    Community cross-list

    Customer-branded remote URL on PulseMCP, Smithery, Glama, and MCPBundles /skills — full-text search surfaces matter for agent discovery.

  4. 4

    Post-launch refresh

    Growth tier includes telemetry and directory maintenance; fix rejected listings and stale metadata.

Directory coverage questions

Is directory coverage the same as Shelf GEO?

Shelf GEO is consumer stores specifically. Directory coverage is the full matrix — stores plus developer and community indexes. Managed MCP Apps includes both.

Why list on MCPBundles /skills if we have our own domain?

Google long-tail (“{brand} mcp”), FAQ schema, and Hub connect paths drive developer discovery. We eat our own cooking on every managed launch.

Does registry search find us by description?

Official registry search matches server name first — encode category and geo intent in the published name (e.g. travelsupermarket-holidays-mcp), not a generic acronym.

What if we only care about ChatGPT?

Launch covers consumer Tier A. Skipping Cursor and Registry leaves IDE agents on a competitor’s endpoint — we scope honestly in intake.

Ready for a directory launch?

Fill out the intake form — we reply within one business day to confirm fit and recommend a package.

Request a directory launch
MCP Directory Coverage — Cursor, Registry, /skills & Community Launch | MCPBundles