com.mcparmory/runpod

Runpod

Is the com.mcparmory/runpod MCP server down? No. It answered initialize and tools/list in the probe of 2026-11-14.

Launch, scale, and manage GPU pods and serverless endpoints across regions

alive probed 2026-11-14

35tools

22ktokens in Claude Opus 5.5

11.1 %of Claude's 200k window

8.8×the median server

8.1 sto initialize

483 msto list tools

Run it: uvx mcparmory-runpod@1.0.1 (stdio, probed in a container with the network off)
Host: pypi.org
Reports itself as: Runpod 4.2.0
Capabilities: logging, prompts, resources, tools
Instructions: none
Of a 1M window: 2.2 % of a 1M context window
Count: 22k tokens in Claude Opus 5.5, exact, measured 2026-10-17. With the Claude Code prefix mcp__…__ on each tool name: 22k. The o200k_base estimate is 12k. Counted again when the tool definitions change. Method.
Repository: https://github.com/mcparmory/registry
Registry: active, updated 2026-05-12 · registry entry

Most expensive tools

The Claude column scales the o200k estimate of each tool by this server's ratio of exact to estimate, 1.83×.

Tool	≈ Claude tokens	o200k estimate	Share of this server
create_pod	2 989	1 629	13 %
update_endpoint_async	2 191	1 194	10 %
create_serverless_endpoint	2 158	1 176	10 %
update_endpoint	2 084	1 136	9 %
list_endpoint_billing_history	1 807	985	8 %

Other registry entries on pypi.org, the ones next to this name; all of them.

Servers that mention runpod.

The result of the latest probe, the tool count and the token cost. The feed has an entry for every probe that changed them.

Stops answering, comes back, tool set or cost changes. One confirmation mail first, a stop link in every mail.

This server is in the heaviest 10 % of the registry. Its 5 most expensive tools are 51 % of the cost. Shorter descriptions, fewer optional parameters and fewer tools cut the cost. The command npx mcpcost --url <endpoint> --tools 50 lists every tool with its cost.