--- title: "PULSAR" description: "Platform for Unified Log Search, Alerting & Review. Free, open-source M365 audit log ingestion with indefinite retention and AI-ready query interface." eyebrow: "Free & Open Source" lead: "PULSAR continuously ingests Microsoft 365 admin audit events and stores them in a database you own. Search, alert, forward to your SIEM, and query via AI — with no retention expiry." icon: "📡" badge: text: "Free & Open Source" color: "green" actions: - label: "View on GitHub" url: "https://github.com/cqrenet/pulsar" external: true primary: true - label: "Learn about ASTRAL →" url: "/products/astral/" --- ## The Problem PULSAR Solves Microsoft 365 E3 provides 90 days of audit log retention, accessible only through the Microsoft Purview portal. Events older than 90 days are permanently gone. There is no alerting on admin actions, no forwarding, and no integration surface beyond Microsoft's API. PULSAR solves this entirely. Once deployed, audit events are continuously pulled and stored in your own MongoDB instance. Retention is indefinite. The data is yours. ## What PULSAR Does
Polls Microsoft audit APIs on a configurable interval and stores all events incrementally. Watermarks ensure nothing is missed. Sources: Entra directory audit logs, Intune audit logs, Exchange / SharePoint / Teams via Office 365 Management Activity API.
A web interface for searching events by workload, operation type, user, resource, and time range. Designed for the kind of investigation that is otherwise done by manually clicking through the Purview portal.
Rules-based alerting with webhook delivery. Trigger on new permanent Global Admin assignments, CA policy changes outside business hours, guest invitations in sensitive groups, and more.
Forward audit events to an external SIEM. Events are normalised and enriched before forwarding — higher-quality data than a direct Microsoft API integration delivers.
PULSAR exposes search_events, get_event, and get_summary via the Model Context Protocol, so an AI assistant (Claude, Copilot, or any MCP-compatible client) can query your audit log in natural language.
Full deployment documentation and Docker images are available on GitHub.