WEBVTT

NOTE Prepared narration track. Final cue timing requires genuine footage.

1
00:00:00.000 --> 00:00:10.000
MCP Queen helps developers find MCP servers, inspect current operational evidence and Trust Receipts, and decide what to connect.

2
00:00:10.000 --> 00:00:30.000
This is the MCP Queen plugin connected to its universal, no-auth production endpoint in ChatGPT Developer Mode.

3
00:00:30.000 --> 00:01:10.000
MCP Queen searches by the developer's task and returns matching servers with endpoints, operational grades, dated observations, and caveats.

4
00:01:10.000 --> 00:01:42.000
The grade breakdown exposes the observations behind each point. It measures observable operational behavior and is not a security certificate.

5
00:01:42.000 --> 00:02:20.000
Tool-level search identifies the specific capabilities developers need and keeps each result connected to its server, endpoint, and evidence.

6
00:02:20.000 --> 00:03:02.000
Trust Receipts keep operational observations separate from security, data integrity, citations, claims, benchmarks, and field evidence. Missing evidence is labeled unaudited.

7
00:03:02.000 --> 00:03:38.000
Existing field reports come from real server use and are human reviewed. They are published as qualitative evidence, never counted as votes, and never directly change the operational grade.

8
00:03:38.000 --> 00:04:08.000
The leaderboard is a current operational comparison, not a universal safety ranking.

9
00:04:08.000 --> 00:04:35.000
MCP Queen rejects feedback based only on search results. A report must describe actual use and enters a quarantined human-review queue.

10
00:04:35.000 --> 00:05:00.000
The same evidence-backed discovery workflow is available in ChatGPT on iOS.

11
00:05:00.000 --> 00:05:25.000
On Android, the Trust Receipt preserves the same distinction between observed evidence and unaudited dimensions.

12
00:05:25.000 --> 00:05:35.000
MCP Queen is the evidence layer for MCP. Find, verify, then connect directly to the server.