Account Start a project
What we build

Custom AI and software for your product or operations, from a first prototype to a system your team runs.

Start a project
How a project runsDescribe, plan, build, hand over

Four steps, a written plan and a fixed price within days.
Subscriptions

Ready-made workspaces by monthly subscription. Try free, cancel any time from your account.

See plans
No account firstPay, and you are in

Your account is created when you subscribe.
Open research

Reports, guides and methods on how AI can fail, free to cite.

All publications
Latest · 10 Oct 2026Coding agents that deleted production data

Two reported cases, in July 2025 and April 2026, in which an AI coding agent destroyed live data, and what they teach about permissions and backups.
Live AI risk

Automated probes, outage data and public reports, reviewed by a person.

Open the monitor
AI risk nowLevel 1 · Calm

6 systems, 2 regions, 0 reports in 24 h.
Before it fails

What to do in the first 72 hours if AI-dependent services fail.

Read the guides
Offline kitSave it now

Guides and data to keep when the internet is down.
Artificial Intelligence Center

We build AI that works, and keep watch on how it can fail.

Custom AI and software for clients, ready-made products by subscription, and an open monitor of AI risk with practical Day Zero guides.

See what we build Open the monitor
Our services what we do
4

Ways to work with us

Custom AI, custom software, AI safety review and ready-made products.

5Workspaces
7Trial days
4Steps
Workspaces by monthly price
  • OSINT monitoring
  • Medical data integration
  • AI marketing team
  • Financial markets monitoring
  • Event-to-market impact
What you keep
100%

Custom work is handed over complete, with training if you want it.
How a project runs
4steps

What we do

All services
Build

Custom AI development

Assistants, agents and data systems built around your own data and tools, from prototype to hand-over.

AI development Build

Custom software development

Web and mobile applications, internal tools, APIs and integrations that are small, secure and maintainable.

Software development Subscribe

Ready-made products

Products we run ourselves, by monthly subscription, with a free trial where it makes sense.

See plans Advise

AI safety review

An outside view of the risks in the AI systems you build or buy, with findings ranked by risk.

AI safety review Prepare

Day Zero

What to do in the first 72 hours if AI-dependent services fail, and what to download while you still can.

Get ready

Latest research

All publications

When agents leave the sandbox: what was reported in 2026

Four reports from March to September 2026 of AI systems leaving the environment they were put in, what they have in common and what is still unknown.
The continuity archive: what to save and howMonitoring method v2: canaries, probes and reportsDo AI models resist shutdown? The evidence so far

How we work on AI risk

Disclosure policy
01

Kept apart from client work.

Nobody can pay us to change a status, a level or a published finding. What we sell and what we publish about AI risk are separate.

02

Open methods.

Test suites, scoring rules and monitor data are published so anyone can check our work.

03

Responsible disclosure.

Developers hear from us before details that could be misused are published. Reporters stay anonymous.

Seen an AI system behave strangely?

Tell us. Reports are anonymous, checked against our probes within minutes and reviewed by a researcher.

Report AI behavior