Your account runs on one pool of RAM, and now that pool is elastic.

Add a RAM pack the moment a build needs more memory, and release it when the spike passes. Each pack is +1 GB with CPU scaling alongside it, priced at EUR 0.16/day per GB and billed by the hour — so a pack you spin up for an afternoon is billed for that afternoon. Extra project slots work the same way.

This is the model the whole platform is built for: capacity that follows your work instead of a plan you size up front. Stack as many packs as a project needs during a heavy build, then let it settle back down. You watch it live in the console, and your agent can resize the pool for you in the same conversation it deploys from.

And because elastic capacity only helps if the bill stays in your hands, one number keeps it there: set a monthly budget, and your agent scales RAM and projects on its own underneath it. Reach the ceiling and new scaling waits for your go, while everything already running keeps running. Raise it, lower it, or go unlimited whenever you like.

See how it adds up on the pricing page.

Today nyxory enters open beta.

The idea behind it is simple: agents already write the code, so they should run it too. nyxory is the runtime for exactly that.

Agent to agent

The part we’re proudest of is the part you never have to touch. nyxory connects to your coding agent as a single MCP server, and from there it’s agent to agent: your agent hands over a repo and a sentence of intent, and the nyxory agent takes it the rest of the way. It picks the build, ships it to a live URL with automatic TLS, and keeps it running — logs, health, restarts, the whole operational loop. The thinking stays with you; the operations move to us, and control comes straight back the moment you want it.

That handover is the whole product. You describe what you want in plain language, agent to agent, and the work happens on our side.

What it does in this release

One hybrid agent, any workload that runs in a container:

• Deploy — hand over a repo; it picks the build and ships it to a live URL with automatic TLS.
• Scale — grow an app’s resources on demand; stack RAM as the work grows, and CPU follows.
• Observe — live status, deploy history, and logs on request, so you diagnose and re-ship in the same conversation.

Beta means we build it with you

The core loop is solid and live, and from here we ship improvements continuously. Your first project runs on the Free tier — see pricing when your work grows. Tell us what breaks or what’s missing; this page is where you’ll watch it land.

Follow along by RSS, or read it as an agent at /changelog.md.