async/await for LLM agents: spawn and await sub-agents scheduled on your own GPUs
Connect Runs on your machine
claude mcp add luftpause -e Runtime__Model=<Runtime__Model> -e DataPlane__Url=<DataPlane__Url> -e LUFTPAUSE_TOKEN=<LUFTPAUSE_TOKEN> -- docker run -i --rm ghcr.io/luftpause/runtime:v0.1.4
{
"mcpServers": {
"luftpause": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"ghcr.io/luftpause/runtime:v0.1.4"
],
"env": {
"Runtime__Model": "<Runtime__Model>",
"DataPlane__Url": "<DataPlane__Url>",
"LUFTPAUSE_TOKEN": "<LUFTPAUSE_TOKEN>"
}
}
}
}
{
"mcpServers": {
"luftpause": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"ghcr.io/luftpause/runtime:v0.1.4"
],
"env": {
"Runtime__Model": "<Runtime__Model>",
"DataPlane__Url": "<DataPlane__Url>",
"LUFTPAUSE_TOKEN": "<LUFTPAUSE_TOKEN>"
}
}
}
}
{
"servers": {
"luftpause": {
"type": "stdio",
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"ghcr.io/luftpause/runtime:v0.1.4"
],
"env": {
"Runtime__Model": "<Runtime__Model>",
"DataPlane__Url": "<DataPlane__Url>",
"LUFTPAUSE_TOKEN": "<LUFTPAUSE_TOKEN>"
}
}
}
}
Replace the values in angle brackets with your own. Keys stay on your machine; we never see them.
Tools
This server runs locally, so its tools are listed by your MCP client once it starts.
Environment variables
| Name | Description |
|---|---|
Runtime__ModelRequired | Model the sub-agents run on, as served by an engine behind the luftpause dataplane |
DataPlane__UrlOptional | URL of the luftpause dataplane that places each turn on the engine pool |
LUFTPAUSE_TOKENOptional, secret | Bearer token required on the MCP endpoint and sent to the dataplane; unset leaves both open |
Details
| Package | ghcr.io/luftpause/runtime:v0.1.4 (OCI, streamable-http) |
| First published | 24 Oct 2026 |
| Last updated | 24 Oct 2026 |
| Registry name | io.github.vmedvedovskiy/luftpause |
More ai and media servers See all
Listing data from the official MCP registry. Status and tools from NestDaddy's daily check. Browse all MCP servers.
