Stop paying for idle databases
Postgres with replication between regions and built-in full-text search, plus a message queue, real-time MQTT messaging and a secrets store. One API request creates each service. An idle database sleeps and uses only storage until the next connection.
$ curl -X PUT \ https://api.northoryx.com/v1/services/postgres \ -d '{"name":"orders-db", "options":{"availability":"sleeping", "maxInstances":3, "regions":["eu-central","us-east"]}}' { "created": true, "service": { "type": "postgres", "name": "orders-db", "regions": ["eu-central", "us-east"] } }
Services
Every backend piece in one account.
Each service is created the same way, through the same API.
Postgres
Managed Postgres that can run in several regions, with replication between them. Full-text search is built in. Data is archived to object storage while the database runs, so a database that sleeps keeps its data.
Queue
Topics, consumer groups, priorities and dead-letter topics. Producers publish over HTTPS; workers via API runtime fetch messages, process them and acknowledge them.
Real-time messaging
An MQTT broker over TCP or WebSocket, for messages between your users' devices and your services, for example chat or live updates.
Vault
Secrets and credentials, encrypted at rest and rotated on a schedule. Encryption uses a key Northoryx manages or a key in your own AWS KMS account.
API runtime
Your code runs as an HTTP API or as a worker for queue jobs, with environment variables you set. In testing, not yet open to early access.
Same model for every service
PUT /v1/services/{type} creates or changes it. Another request deletes it. Upgrades and failover are done by us.
Scale to zero or keep it running
No instance sizes or tiers to pick.
Databases add instances as load rises and remove them as load falls. You choose, per database, what happens when nobody is using it. The queue, real-time messaging and secrets run on shared clusters that are always up.
The database sleeps when idle and wakes on the next connection. While it sleeps, its data is kept in object storage and no instance runs.
The problems it solves
Production services without an operations team.
- 01
You don’t want to run infrastructure.
Running Kubernetes, databases, backups and upgrades yourself takes engineers away from your product. On Northoryx Cloud there is nothing for you to operate. You call the API and use the endpoint.
- 02
You pay for services that sit idle.
Staging, preview branches, internal tools and side projects run all day and night on most clouds. A Northoryx database sleeps when idle and starts again on the next connection. The queue, messaging and secrets run on shared clusters, so an idle app keeps no servers of its own running.
- 03
“Serverless” databases still charge a minimum
Using “serverless” databases or search engines you often find yourself with much bigger bills than expected. Real price is often hidden from you, highly complex multi-region setups, that sometimes require approvals.
- 04
Managed services are expensive.
Northoryx Cloud is built to run partly on AWS spot instances and Hetzner servers, which cost less than on-demand cloud servers. A database copies every finished WAL segment to object storage while it runs, so when a machine is taken away, the database starts on another machine from that copy.
- 05
Every backend piece is a separate vendor.
Postgres from one provider, search from another, a queue from a third, secrets from a fourth. Here every service is in one account and created the same way. Full-text search is part of Postgres.
- 06
Traffic between clouds is hard to secure.
Our servers talk to each other over encrypted WireGuard tunnels, across every cloud and data center we run on. You do not set anything up.
- 07
You have to guess how big your servers should be.
Too small and a spike stops your service; too large and you pay for capacity you don’t use. Northoryx adds and removes instances and servers as load changes, with no action from you.
What you get
We run it, scale it, back it up and upgrade it.
One API. One request creates or changes a service; another deletes it.
Autoscaling. Instances and servers are added as load rises and removed as it falls.
Wakes on connection. A sleeping database starts when a connection arrives. The connection is held until the database is ready.
Postgres in several regions. A database can run in more than one region, with replication between them. When two regions write the same row, the later commit wins, on every table.
Data archived while it runs. Every finished WAL segment is copied to object storage. If a machine fails, at most the segment being written is lost.
Nothing to operate. Upgrades, failover and data archiving are done by us.
Your own encryption key. Secrets can be encrypted with a key in your own AWS KMS account instead of ours.
Encrypted traffic between servers. WireGuard across every provider we run on, with no VPN setup on your side.
Estonian company. Northoryx is built and run by a company registered in Estonia, in the EU.
Who it is for
Northoryx Cloud is live in 2027.
Get notified when we launch. No spam. One email when Northoryx is ready.
Already have early access? Sign in