Independent research lab · Barcelona · est. 2026

We build operating systems for autonomous software.

Agents today are pets: hand-configured, fragile, impossible to reproduce. We think they should be infrastructure: declared, versioned, instantiated on demand. We build the machine that treats them that way.

scroll

01Premise

Autonomous software is an operating-system problem

Every team running agents today rebuilds the same fragile scaffolding by hand. Agents belong in the same category as the rest of your infrastructure, and putting them there is a systems problem nobody has finished solving.

01

Every agent is a pet

Hand-configured, stateful and fragile. You can't reproduce one, you can't say what it will touch, and when it breaks there is nothing to roll back to.

02

Patching the wrong layer

Guardrails, approval prompts, tool allow-lists, all written in the same process the agent controls. A boundary only counts if the thing inside can't reach it.

03

Proven parts, no system

Identity, isolation, supervision, an audit trail. Proven for decades, never assembled into a system an autonomous process can run on. That gap is the opportunity.

02Primitives

Six primitives the operating system already provides

Each one is settled at the systems layer, with decades of production use behind it. None of them were designed for software that acts on its own, and making them hold for that workload is the work.

Identity

Agents get a system account of their own, separate from yours.

a user

Environment

Its tools are a declared, reproducible set. Nothing else exists.

a build closure

Authority

What it may read, write and reach is enforced by the kernel.

a sandbox

Execution

Each task is a supervised process, created and removed on demand.

a transient unit

Memory

Durable state, plus an append-only record of everything that ran.

a store and a journal

Agency

The model is a component inside it, not the architecture.

a runtime

The mechanisms are proven. What does not exist yet is the system that holds them together for software that acts on its own, and that is what we are building.

03The flagship

PRAXIS

An operating system for autonomous software.

Praxis is the first system we've built on those primitives. You declare what a kind of agent is; the machine gives every one of them a real identity, a fixed set of tools and a boundary it cannot argue its way out of.

declared and versioned
  1. 01 Class what a kind of agent is
created on demand
  1. 02 Instance one agent, with its own memory
  2. 03 Task a piece of work you hand it
  3. 04 Execution one supervised process

Keeping these four apart is the whole design.

praxis

          

The control surface is commands, not conversation.

  1. runcreate an instance
  2. taskgive it more work
  3. resumepick it up again
  4. forkbranch its work
  5. stopend the execution
  6. deleteerase its state too

Agents are objects

They exist when nothing is running. An agent is state on disk, not a live process.

Tools are capabilities

An agent's toolset is declared up front. What you didn't give it is absent.

Boundaries are enforced

Isolation comes from the kernel, so a boundary holds whatever the model tries.

Everything is inspectable

One command prints exactly what an agent may do, and what enforces each line.

04The lab

A lab that builds what it researches

Drac Labs is an independent research lab working on how autonomous software is identified, contained, observed and stopped.

Everything we learn goes into what we ship, and everything that breaks goes back into the research. It's a short loop, and it's the whole method.

What we work on

Operating systems for software that acts on its own
Capability and isolation models for agents
Runtimes, durable state and the audit trail
01

Foundations can't be retrofitted

Identity, isolation and audit either exist in the design or they never arrive. So that is where we start.

02

Only claim what something enforces

A tool that overstates its guarantees is worse than one that promises nothing, because you'll trust it.

03

Enforcement, not persuasion

Anything you can talk your way past isn't a boundary. We work where the model has no vote.

04

Build for the decade

Operating systems are measured in decades. That horizon decides what we build and what we leave alone.

05Draco

The dragon was always in the sky

Draco winds between the Bears and never sets: a circumpolar constellation that has been overhead every night of recorded history. Four and a half thousand years ago one of its stars, Thuban, was the pole star: the fixed point everything else turned around.

We took the name because it is the right metaphor for the work. Something enormous, old, and useful precisely because you can navigate by it, for as long as you keep watching it.

The figure drawn in the background is the real thing: seventeen stars at their true right ascension and declination, stereographically projected around the asterism's centroid.

StarBayerMagNote
Thubanα Dra3.65pole star, c. 2700 BCE
Eltaninγ Dra2.24brightest in Draco
Rastabanβ Dra2.79"head of the serpent"
Athebyneη Dra2.73orange giant
Altaisδ Dra3.07the neck bend
Aldhibahζ Dra3.17blue-white B-type
Edasichι Dra3.29first giant with a planet
Scroll to draw the asterism ↑

06Contact

Come and see the dragon.

We're early, and we'd rather talk to a handful of the right people than announce anything. If one of these is you, we should speak.

01

Investors

Deep tech and infrastructure, with a decade-long view of where agents will run.

02

Design partners

Teams already running autonomous software somewhere it actually matters.

03

Researchers

Systems people who want to work where the guarantees are actually enforced.