M9K · Data platform

The engine under critical infrastructure.

One compiled platform — objects, metrics, files and AI behind a single API. The foundation our own product, BigBoard, runs on, in production today.

See the platform What's running today
Compiled·dependency-light·hosted or on-prem
The platform

One platform where most teams wire together five.

Object store, time-series, auth, files and an AI layer — one engine, one API, built to work together.

A high-performance metrics engine

Query a decade as fast as an hour — statistics computed as data arrives, every series histogrammed for behavioral change, not just current value.

flat query response · pre-aggregated · group-by & rollup

Object store

Model anything; add a type without a schema migration.

CRUD · import/export · dependencies

Auth & RBAC

OAuth/JWT, scoped keys, per-object permissions.

per-object roles

Server-side AI

Data-aware channels; the API key never reaches the browser.

streaming · M9K-aware

Search

Full-text plus attribute, type and metric-range filters.

BM25 · type wildcards

Live events

Object, metric and dependency changes pushed over WebSocket.

wss · object-bound

OpenTelemetry-native

OTLP metrics land directly, cardinality handled server-side.

/v1/metrics

Versioned files

Attach files to any object; history and range downloads.

resumable

Defense at the edge

Injection & scanner detection, per-IP threat scoring, rate limits.

auto-block
Why it's built this way

Lean by construction.

01

Few dependencies

In-house JSON & HTTP, no npm or PyPI package tree — almost none of the supply-chain surface most breaches ride in through.

02

Efficiency is the economics

As the AI buildout bids up compute, memory and power, a compiled core on dedicated hardware spends a fraction of a cloud stack.

03

A footprint you can watch

Native per-OS agents — a few hundred kilobytes — that report their own CPU and memory, so the footprint is measured, not asserted.

Proof, not promises

It's not a deck. It's running.

13
Monitoring agent types, shipped
0
npm / PyPI packages — no tree to poison
1hr=10yr
Same metric query time, measured on the live API
1
Full product (BigBoard) in production on the platform

Agents span FreeBSD, Linux, macOS, Windows, SNMP, UPS, PDU, Modbus, gNMI, ICMP, syslog and external market feeds — all reporting through one platform.

Architecture

One API. Many tenants. No package tree.

  • Multi-tenant within the APIEvery call is namespaced by instance; each tenant's data and access stay separate inside the platform — not a server per customer.
  • Compiled native servicesNo interpreter, no heavy runtime — far less hardware to do the same work.
  • Hosted, or on your premisesRun it on infrastructure we operate, or have us deploy and manage a sealed appliance in your own facility — no hyperscaler between you and your data.
m9k_api.py
# one wrapper, the whole platform
from m9k_api import api

api('POST', 'object', {'parent_id':'0',
  'content':'rack-7', 'type':'device:rack'})

# ten years or one hour — same query, same speed
api('GET', 'metrics/cpu?group=site&from=-10y')
The first product on M9K

BigBoard — ask your critical infrastructure

Observability as a business tool: ask in plain language, get an answer grounded in live data — for engineers, field technicians, and the people deciding what to spend. Proof the platform carries a real product, not just an API.

Visit BigBoard →
Deployment

Hosted by us, or on your premises.

Hosted

We run the platform on infrastructure we operate. You build and use it through the API.

  • Dedicated hardware, not rented cloud
  • Predictable cost as you grow
  • Operated by the people who built it

On-premises appliance

When data has to stay in your facility, we deploy and operate a sealed M9K appliance on-site.

  • Your data never leaves your premises
  • We keep it patched, monitored and running
  • Same API as the hosted platform
Company

A small team that ships — two founders doing the engineering and the operations, choosing depth over headcount.

Pre-launch by design: the platform and its first product are proven in production, and we're opening to early partners and investors who prefer to evaluate software that already works.

Talk to us →