Developers & API
MCP

Your monitors, available to AI clients — under policy.

PulseTrace speaks the Model Context Protocol. Register an AI client, grant it scopes, and it can query monitors, incidents, posture and Intelligence — with each tool call policy-gated: some allowed outright, others requiring approval, and every call and session logged.

PulseTrace runs a first-class MCP server whose tools span monitors, executions, security posture, Intelligence and notifications. You register a client, grant it scopes, and each tool call is evaluated against your policy — allowed outright, or held for an explicit approval — with every call and session written to an activity log. Access is always least-privilege and fully auditable, so an agent can be useful without being trusted blindly.

  • Tools across the platform — monitors, executions, posture, intelligence, notifications.
  • Scoped, policy-gated access — allow a tool outright, or require an approval.
  • Every call & session logged under MCP → Activity.
mcp · connected clients
claude-desktop
scopes: monitors:read · incidents:read
Active
ops-copilot
tool call · list_incidents(status=open)
Policy-gated
runbook-agent
requires approval to acknowledge
Approved