nano SIEM

Introduction to nano

Introduction to nano

nano is an AI-powered SIEM for security teams that need fast log analytics.

Whether you're ingesting 2 GB/day on Hobby or running unlimited volume on Enterprise, nano covers the full workflow: detection, alerting, investigation, and response. Events are stored in ClickHouse for fast queries, with pivt AI for parser, query, and detection assistance.

Why nano?

Fast

  • Sub-second queries across billions of events using ClickHouse
  • Real-time detection with sub-30s latency via materialized views
  • High-throughput ingestion with Vector: HTTP, Syslog, HEC, and native Vector protocol

AI-Powered

  • pivt AI generates parsers, queries, detection rules, and investigation summaries from natural language
  • Auto-tuning reduces false positives by analyzing historical detection matches
  • Shadow Investigation autonomously hunts for threats when cases are created
  • Multi-provider: Anthropic, OpenAI, and Google, routed through Cloudflare AI Gateway

Query Language

  • nPL (nano Pipe Language): piped syntax, e.g. source_type="cloudtrail" | stats count by user | sort -count
  • 100+ eval functions, including security-specific ones like cidrmatch, ispublicip, defang, and entropy
  • Advanced analytics with timecharts, window functions, and statistical analysis

Security-First

  • Risk-based alerting with per-entity risk scoring
  • Prevalence tracking: flag rare and first-seen indicators automatically
  • MITRE ATT&CK mapping for all detection rules
  • Unified Data Model: 505 normalized fields across 26 categories for cross-source correlation
  • Case management with AI-powered investigation timelines

Architecture

Key Components

ComponentTechnologyPurpose
Log CollectorVectorHigh-performance ingestion, parsing, and routing
Log StorageClickHouseTime-series log storage with sub-second analytics
Metadata StorePostgreSQLRules, alerts, dashboards, users, configuration
APIRust (Axum)Management API for rules, alerts, settings, ingestion
SearchRust (Axum)Query execution, field stats, async jobs
FrontendReactWeb interface with lazy-loaded components
AI Assistantpivt (via Cloudflare AI Gateway)Parser creation, query building, detection tuning, investigation
MarketplaceBuilt-inEnrichments, threat intel feeds, and community integrations

Deployment Options

nano deploys as a managed service or into your own cloud (BYOC):

TierCapacityInfrastructure
Hobby / Startup2-5 GB/daySingle VPS (Docker Compose)
Growth10 GB/dayKubernetes (single pool)
Team / Business25-50 GB/dayKubernetes with HA
Pro100 GB/dayMulti-pool Kubernetes
EnterpriseUnlimitedClickHouse Cloud + managed PostgreSQL

See Deployment Architecture for full details on each tier.

Get Started

  1. Set Up Your First Feed: import a pre-built parser or create one with the AI wizard
  2. Search Your Data: learn nPL to hunt through your logs
  3. Create Detection Rules: automate threat detection
  4. Configure pivt AI: set up providers, models, and organizational guidance
  5. Deploy On-Prem Collection: collect Windows, Linux, and network logs
On this page

On this page