OpenApps

OpenApps answers questions about the app stores inside AI assistants: what is listed, ranked, featured and launched in the your AI, your AI, Cursor, Grok Bot, Perplexity and Muse stores, and which tools their apps' MCP servers expose.

Ask for:

What you can ask

Review a tracked MCP server's tools (names, descriptions, annotations, schemas) against patterns across every open server OpenApps sees.
Example path →

MCP connector

Server URL: https://theopenapps.com/mcp/claude
Login: API key
Protocol: MCP 2026-07-28
Instructions: OpenApps tracks the ChatGPT (per country), Claude, Cursor (with cursor.directory), Grok Bot, Perplexity and Muse app stores, plus the MCP servers their apps link to and those servers' tools. Not for phone app stores, install or revenue estimates, or submitting apps. Conventions: a listing ref is '{store}:{plugin_id}' ( finds it; an OpenApps URL works too). Times are ISO-8601 or relative to the latest data (-24h, -7d, -2w); windows default to 24h. region is a ChatGPT country code (default us); the other stores have one region each. lists each store's charts, kinds and metrics. limit defaults to 20 rows; to page, pass the answer's next_cursor as cursor. Answers are text: a header line, CSV rows, then next_cursor, truncation, plan and coverage lines where they apply. response_format=detailed adds structuredContent, the full { data, meta } record with every field. Signed out: the United States and a list's top 200 rows; signed in, a free account adds every country and 1,000 rows, Pro all (the plan line says what was left out). Word claims by their basis: observed (seen in the store), store_reported (published by the store, e.g. Claude installs), derived (computed by OpenApps), proxy (e.g. GitHub stars) or estimate. Say when the data is from (the header's as_of or data-from time). An app absent from a country the store refuses (coverage unavailable) is not "not listed". Rows carry a url to the OpenApps page. A next line names a call that usually follows; for a brief, a comparison or a server audit, guide=<playbook> gives the steps.

Tools

Reads data11

Deletes or can't be undone1

Skills

None in the store listings or on the server

What you can ask

Review a tracked MCP server's tools (names, descriptions, annotations, schemas) against patterns across every open server OpenApps sees.

Example path →

MCP connector

Server URL

https://theopenapps.com/mcp/claude

API key

Protocol

MCP 2026-07-28

Instructions

OpenApps tracks the ChatGPT (per country), Claude, Cursor (with cursor.directory), Grok Bot, Perplexity and Muse app stores, plus the MCP servers their apps link to and those servers' tools. Not for phone app stores, install or revenue estimates, or submitting apps. Conventions: a listing ref is '{store}:{plugin_id}' ( finds it; an OpenApps URL works too). Times are ISO-8601 or relative to the latest data (-24h, -7d, -2w); windows default to 24h. region is a ChatGPT country code (default us); the other stores have one region each. lists each store's charts, kinds and metrics. limit defaults to 20 rows; to page, pass the answer's next_cursor as cursor. Answers are text: a header line, CSV rows, then next_cursor, truncation, plan and coverage lines where they apply. response_format=detailed adds structuredContent, the full { data, meta } record with every field. Signed out: the United States and a list's top 200 rows; signed in, a free account adds every country and 1,000 rows, Pro all (the plan line says what was left out). Word claims by their basis: observed (seen in the store), store_reported (published by the store, e.g. Claude installs), derived (computed by OpenApps), proxy (e.g. GitHub stars) or estimate. Say when the data is from (the header's as_of or data-from time). An app absent from a country the store refuses (coverage unavailable) is not "not listed". Rows carry a url to the OpenApps page. A next line names a call that usually follows; for a brief, a comparison or a server audit, guide=<playbook> gives the steps.