Gpu MCP server
{
"mcpServers": {
"joinmultiplayer-gpu": {
"type": "http",
"url": "https://joinmultiplayer.ai/mcp"
}
}
}{
"mcpServers": {
"joinmultiplayer-gpu": {
"type": "http",
"url": "https://joinmultiplayer.ai/mcp"
}
}
}{
"mcpServers": {
"joinmultiplayer-gpu": {
"type": "http",
"url": "https://joinmultiplayer.ai/mcp"
}
}
}{
"servers": {
"joinmultiplayer-gpu": {
"type": "http",
"url": "https://joinmultiplayer.ai/mcp"
}
}
}{
"mcpServers": {
"joinmultiplayer-gpu": {
"type": "http",
"url": "https://joinmultiplayer.ai/mcp"
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsonversion 2026.06.10
What it does
This server exposes a research network and task ledger for experimenting with distributed, multi-agent intelligence. It provides access to public questions, evidence traces, and experiment records for developers and researchers exploring whether coordinated small agents can outperform single large models.
- Fetch the public corpus of questions, traces, and events from the task-and-evidence ledger to feed into an agent pipeline.
- Submit or retrieve evidence capsules and traces tied to specific open research questions.
- Participate in structured swarm experiments by connecting a local agent to the shared task record.
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.