GuardRails-MCP runs on your own machine — the client starts it, so there is no endpoint to ping. 241 installs a week from pypi. Last commit 16 Oct 2026.
Local prompt regression tracker with six fictional cases and evidence for human review.
Today is the operative word: we check GuardRails-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 guardrails-mcp -- uvx guardrails-mcp
{
"mcpServers": {
"guardrails-mcp": {
"args": [
"guardrails-mcp"
],
"command": "uvx"
}
}
}
[mcp_servers.guardrails-mcp]
command = "uvx"
args = ["guardrails-mcp"]
{
"mcpServers": {
"guardrails-mcp": {
"args": [
"guardrails-mcp"
],
"command": "uvx"
}
}
}
{
"mcpServers": {
"guardrails-mcp": {
"args": [
"guardrails-mcp"
],
"command": "uvx"
}
}
}
This one needs environment variables set before it will start:
GUARDRAILS_DATA_DIR (Absolute writable directory for local evaluation records. Use the same path in CLI and dashboard.).
The author declared them in the registry entry; get the values from the project itself.
Local read-only MCP scanner for known prompt-injection patterns with exact evidence and no network.
Guarded local Charles Proxy automation and bounded XML session evidence
Regression-test AI models on real Snapdragon and Jetson hardware, with signed evidence for CI.
Transactional refactoring MCP with 11 evidence-backed tools, hash-bound previews, and rollback.
Local-first TDLib Telegram MCP server with 115 tools, account isolation and production guardrails.
Verified fixes, professional reviews, and evidence-led apparel over MCP.
Local-first, source-cited memory for AI agents with proposal review, failure memory, and MCP tools.
Transactional refactoring extension with 11 evidence-backed tools, previews, and rollback.
Answers built from our own checks of this server.