MCP Station Report: Remote Server Changes, November 2, 2026

Reachability

MCP Station’s daily check for November 2, 2026 covered 36,456 servers in the index. Of these, 32,004 were reachable as remote servers, a share of 88 percent. Between November 1 and November 2, 139 servers came back after being unreachable, and 171 went down. The net movement was 32 more servers going down than coming back.

For people and agents that connect to MCP servers, a server that stops answering means failed tool calls or an empty tool list at connection time. A server that comes back may no longer offer what it did before, so a client that cached its tool list may need to refresh it.

Listings and tool changes

In the same two-day period, 1,148 servers were newly listed. Tool changes were recorded as follows:

• 355 servers added tools, for a total of 1,284 tools added.
• 99 servers removed tools, for a total of 538 tools removed.
• 524 servers changed a schema, affecting 1,884 tools.

Tool additions widen what an agent can do, but an agent that discovers tools dynamically will see a different set than it saw before. Removed tools can break workflows that name them directly. Schema changes matter because they alter the inputs or outputs a tool expects. A call that worked yesterday may be rejected today, or may return data in a different shape. Schema changes affected more tools than additions and removals combined, so clients that validate arguments against a stored schema have a clear reason to re-check it.

New destructive tools

The index recorded 112 new tools flagged as destructive, spread across 42 servers. A destructive tool can delete or overwrite data or cause other changes that are hard to reverse. When such tools appear on a server an agent already uses, approval settings, permission scopes and confirmation steps may need review. Agents that auto-approve every tool from a connected server would extend that approval to the new tools.

Reading the numbers

These figures are observations from two days of checks. They describe what changed, not why, and they do not say whether any change is an improvement or a problem. Category breakdowns were not available for this report.

Measured changes

32,004 of 36,456 remote MCP servers answered on the 2026-11-02 check.

Largest categories

About this data

Measurements by MCP Station, which checks every remote server in the Official MCP Registry daily. Measurements are not a security guarantee and are not an endorsement of any server. Methods: mcpstation.net/methods.html. Data: mcpstation.net/api.

Disclosure: the operator of Emerge Times also operates MCP Station.

This article was prepared with AI assistance and reviewed by an Emerge Times editor before publication. Licensed CC BY 4.0: credit “Emerge Times (emergetimes.com)”. Usage & License