Vynn MCP runs on your own machine — the client starts it, so there is no endpoint to ping. 127 installs a week from pypi.
Self-improving AI workflows with backtesting, prompt optimization, and scheduling
Today is the operative word: we check Vynn MCP every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.
This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.
claude mcp add vynn-mcp -- uvx vynn-mcp
{
"mcpServers": {
"vynn-mcp": {
"command": "uvx",
"args": [
"vynn-mcp"
]
}
}
}
[mcp_servers.vynn-mcp]
command = "uvx"
args = ["vynn-mcp"]
{
"mcpServers": {
"vynn-mcp": {
"command": "uvx",
"args": [
"vynn-mcp"
]
}
}
}
{
"servers": {
"vynn-mcp": {
"type": "stdio",
"command": "uvx",
"args": [
"vynn-mcp"
]
}
}
}
Multi-model AI orchestration with 31 tools, YAML workflows, and 5 token-optimized profiles.
MCP server for agent task workflows with phases, prompts, gates, and multi-agent coordination
Scan options flow, analyze real-time market data, and build investing and trading workflows.
Automate tasks, processes, and approvals with AI.
Analyze web performance and get optimization insights from GTmetrix, directly in your AI workflow.
Multi-step workflow orchestration for VMware MCP skills with approval gates and rollback.
Describe an automation in plain English and get a working FlowRunner flow with approval steps.
Multi-step workflow orchestration for VMware MCP skills with approval gates and rollback.
Answers built from our own checks of this server.