Real-Time Analytics

Hanzo Database

Real-time analytics database

Sub-second queries over billions of rows. Columnar storage, vectorized execution, and incremental materialized views. Built for product analytics, observability, and customer-facing dashboards.

Columnar
Storage
Sub-sec
Latency
Billions
Rows
Postgres
Wire

Built for Analytics, Not Compromised by It

Time-series, events, metrics, traces — one engine.

Vectorized Execution

SIMD-accelerated columnar engine. Scan billions of rows per second per core. Push-down predicates and projection-only reads.

Incremental Materialized Views

Define rollups, aggregations, and joins as views. Hanzo Database keeps them fresh on every insert. No batch ETL.

High-Cardinality Metrics

Million-tag dimensions without downsampling. Per-user, per-device, per-experiment slicing — interactive at any scale.

Tiered Storage

Hot data on NVMe, warm on SSD, cold on object storage. Transparent migration based on age and access patterns. Pay for what you query.

PostgreSQL Wire

Drop-in for any psql-compatible client, BI tool, or ORM. Tableau, Metabase, Superset, dbt — they all just work.

Zero-Downtime Schema

Add columns, change types, alter partitioning live. Backfill in the background. Online DDL without read-your-write surprises.

Standard SQL. Real-Time Reads.

events.sql
-- Wide event table, partitioned by hour
CREATE TABLE events (
  ts          TIMESTAMPTZ NOT NULL,
  user_id     UUID,
  event       TEXT,
  properties  JSONB
) PARTITION BY RANGE (ts);

-- Always-fresh per-user daily counts
CREATE INCREMENTAL MATERIALIZED VIEW dau AS
SELECT
  date_trunc('day', ts) AS day,
  user_id,
  count(*)              AS events
FROM events
GROUP BY 1, 2;

-- Sub-second over a billion rows
SELECT day, count(distinct user_id) AS dau
FROM dau
WHERE day >= now() - interval '30 days'
GROUP BY 1 ORDER BY 1;

Get started with Database

Open source

License: Apache-2.0hanzoai/database

Get Database

Unified database