You build.We keep it running.

Infrastructure that tunes itself, so your team ships.

Torus watches every server, queue and bill, then fixes what drifts before anyone gets paged.

Book a demo A 30 minute walk through your own stack.

99.98^%: UptimeLast 90 days
41^%: Lower cloud billMedian, first quarter

Torus Core 4Latest release

We run the machines underneath, so you can spend your days on the product above them.

Products

001Torus

Built around one idea: infrastructure should notice problems before people do.

Five tools that share one model of your stack. Use one, or let them work together.

Perforated silver server front, Torus Core

Featured productTorus Core The engine underneath everything. Core learns how your systems behave on a normal day and steers them back when they drift.

Product lineup5/5

One model of your stack, five ways to use it.

Every Torus tool reads the same live map of your services, so a fix in one place is known everywhere.

Shared field model

Flux Router Moves traffic to the healthiest region in milliseconds.

Coil Scaler Adds and removes capacity before the spike arrives.

Pole Monitor One quiet signal instead of a hundred noisy alerts.

Lattice Guard Locks down access and patches as threats change.

Black glossy geometric blocks, Torus Core 5 visual

Torus Core 5.0b

Early access opens this winter.

Build progress

72.4^%

Join the beta

See the whole platform

Services

002Torus

Hands-on help to set up, tune and guard modern infrastructure.

Our engineers work next to yours until the system runs on its own.

Core services4/4

Set up once, tuned for good.

Each service ends with Torus running the work on its own. What you get:

Plan, connect, tune and guard.

See every service

Process

003Torus

Four steps from the first call to a stack that runs itself.

Most teams are live within two weeks.

How it works

Step 1 - Connect

Torus reads your cloud accounts, clusters and pipelines through read-only keys. Nothing changes yet, we only listen.

Step 2 - Map

In a few days Torus learns what a normal Tuesday looks like for every service, queue and database you run.

Step 3 - Tune

We switch on fixes one group at a time. You approve each rule once, then Torus applies it whenever it is needed.

Step 4 - Guard

Lattice Guard watches access and patches around the clock and learns from every attempt it blocks.

See the four steps run on a copy of your stack.

20^min

Average connection time

4^days

Average learning time

38^%

Fewer pages in month one

9^sec

Average threat response

Book a demo

How we work with your team

Analytics

004Torus

Every number that matters, live, in one calm view.

No dashboards to build. Torus ships them ready.

Live Field view

Torus

System health

Across all regions

71%: Optimal
22%: Stable
7%: Watching

AmsterdamUTC

Tuned31ms to 12ms

TorontoUTC

Tuned44ms to 18ms

SingaporeUTC

ScalingAdjusting

RegionsResponse time, live

90-day window

99.98^%

Uptime

Performance Requests per second, by quarter

96%: With Torus
54%: Before

Score by area

42%Setup

76%Speed

94%Security

81%Scaling

98%Uptime

Torus score

88/100

99.4^%

Core capacity

Numbers

Inside the platform

Pricing

005Torus

Clear plans that grow when you do.

Pay for what you run. Change plans any month, nothing locked in.

Starter

Price

$59

Save 16%

Plan

Core

Seats

For small teams

Core monitoring
Auto restarts
Basic security
Email support
7-day history

Choose plan

Growth

Price

$119

Save 16%

Plan

Advanced

Seats

For teams that scale

Everything in Starter
Coil Scaler
Flux Router
Chat support
90-day history
Cost reports

Choose plan

Fleet

Price

$349

Save 16%

Plan

Enterprise

Seats

Unlimited

For large organisations

Everything in Growth
Lattice Guard
Private region
Named engineer
SLA 99.99%
Audit export

Choose plan

Compare every plan

Questions

006Torus

Straight answers, before you book a call.

Still unsure? An engineer replies within a day.

Torus connects to your cloud, learns how your systems behave on a normal day and fixes drift on its own: restarts, scaling, routing and patches.

No. Torus works with the clouds, clusters and pipelines you already run. It connects through read-only keys first.

Connecting takes minutes. Learning a normal week takes a few days. Most teams switch on the first fixes in week two.

You do. Every rule is approved once by your team, and every change is logged with the reason it was made.

Torus reads metrics and logs, never customer records. Everything is encrypted, and Fleet can run in your own region.

Email on Starter, chat on Growth and a named engineer on Fleet. All plans include the onboarding calls.

A flat monthly price per plan with a set number of seats. Yearly billing saves 16 percent.

Yes. Cancel in one click and export everything Torus learned about your stack.

Still have questions?

Every stack is different. Tell us about yours and an engineer will show you how Torus would run it.

Talk to an engineer

All questions

Team

007Torus

The engineers behind Torus. Builders, operators, fixers.

We have all carried the pager. Torus is the tool we wanted then.

Torus team

Ada Varela

Co-founder, CEO

I spent ten years running infrastructure for payment companies, mostly at three in the morning. I started Torus so the next generation of engineers can sleep through the night while their systems quietly fix themselves.

Bastian Castellane

Co-founder, CTO

I lead the model that sits at the heart of Torus Core. My job is to make sure it learns fast, explains every decision it makes and never touches anything it does not fully understand.