Hayward runs on your own machine — the client starts it, so there is no endpoint to ping. 5 695 installs a week from pypi. Last commit 1 Nov 2026.
Checks ML model files (pickle, PyTorch, joblib, NumPy, GGUF, ONNX) for code that runs on load.
Nothing serious here today
Today is the operative word: we check Hayward every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.
Connect this server
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 hayward -- uvx hayward{
"mcpServers": {
"hayward": {
"command": "uvx",
"args": [
"hayward"
]
}
}
}[mcp_servers.hayward]
command = "uvx"
args = ["hayward"]{
"mcpServers": {
"hayward": {
"command": "uvx",
"args": [
"hayward"
]
}
}
}{
"servers": {
"hayward": {
"type": "stdio",
"command": "uvx",
"args": [
"hayward"
]
}
}
}Alternatives to Hayward
same job, measured the same wayScans ML model files (PyTorch, joblib, NumPy, GGUF, TensorFlow…) and HF repos for supply-chain risks
Unofficial MCP server for Google Colab: runtimes, code, long jobs, files and notebooks.
Mac app that runs Claude or Codex on a schedule against your mail, calendar, and files.
One record of code, files and pages that people and agents read and write, with lanes and git.
Runs physverify checks on physics code. It runs the code it checks: use on trusted code only.
Static checks for Meta _fbc, _fbp and fbclid code that cause modified click ID warnings.
Run terminal processes, edit files and search code on the local machine from an MCP client.
Find the LLM/AI calls you didn't need — where plain code or a regex does the job. No model.
Hayward — questions
Answers built from our own checks of this server.