DB
Back homeIndustry

SaaS & infrastructure

Ship analytics, materialized views, streaming, and search from a single small binary you can run anywhere.

The problem

Why this is hard today

SaaS teams reach for a database, then a cache, then a stream processor, then a search index — and end up operating four systems to ship one product feature.

Each addition is more schema duplication, more consistency glue, and more surface to run in every environment from a developer laptop to production.

A single engine that covers OLAP, OLTP, and streaming behind one contract — in a 154MB image — collapses the operational surface without giving up capability.

Where NYXDB fits

Use-case journeys

Product analytics & usage metrics

Turn raw product events into live metric tables with continuous transforms.

Observability & logs

Abuse & fraud prevention

Read-your-writes per-account state and exact counts for real-time abuse checks.

Fraud & risk

Telemetry & infra monitoring

Line-rate ingest of infra telemetry with instant current-state reads.

IoT & telemetry

Architecture

How NYXDB fits SaaS & infrastructure

One modular engine spans append/keyed/attribute tables, continuous transforms, streaming reads, and vector search, driven by SQL and a Protobuf DSL.

  1. 01

    Model

    Append, keyed, and attribute tables in one catalog.

  2. 02

    Derive

    Continuous transforms maintain materialized views.

  3. 03

    Serve

    SELECT, STREAM SELECT, and vector search from one engine.

  4. 04

    Deploy

    One 154MB image from laptop to production.

Real SQL

Representative query

Live usage rollup — transform
CREATE TABLE usage (
  tenant String NOT NULL,
  calls  UInt64,
  PRIMARY KEY (tenant)
) DELTA (keep = 'latest');
CREATE TRANSFORM roll_usage
  INTO usage AS
SELECT tenant, count() AS calls
FROM api_events
GROUP BY tenant;

Every statement follows the engine’s own test SQL shapes. See the SQL reference for full syntax.

Capabilities

What you get

One engine

OLAP, OLTP, and streaming behind one contract.

Materialized views

Continuous transforms keep them fresh.

859+ native functions

Rich SQL plus a Protobuf DSL.

154MB image

Runs the same everywhere.

Proof

Measured where it counts

gap-free

materialized-view recovery

ADR-080

865

native functions

function manifest

154MBTODO-verify

container image

placeholder — not yet substantiated

▲Figures marked TODO-verify are placeholders pending a published, reproducible benchmark; substantiated numbers cite their source.

Learn more

Related documentation

Transforms

Materialized views.

Functions

865 native functions.

Deployment

Run the image.

Fewer systems, same capability

Explore the docs or run a node.

Get started Read the docs