Managed OmniRun

We run the hosts. You run your agents.

The same open-source sandboxes, operated by the people who build them, on hardware in Finland. For teams that want microVM isolation without owning a rack.

Talk to us Or self-host it

Team workspace

Example team · eu-fi-1

MJA

Sandbox-hours this month412.6 h of 600 h agreed quota

claude-code 228.1 hpython-3.11 114.0 hnode-20 70.5 h

CI runners
used 2 min ago
omr_4f2c81…Active
Agent backend
used 14 s ago
omr_9b17e0…Active

3 members · 2 tokensAudit log on

EU hostingAll hosts in Finland

No lock-inApache-2.0, self-host any time

Team controlsTokens, quotas, audit log

Agreed termsLimits and price set with you

What we handle

Everything below the API.

You call the API. We keep the machines, templates and controls around it running.

Hosts and upgrades

Bare-metal hosts with KVM, kernel and Firecracker updates, storage, monitoring and incident response. You never see a server.

Templates

python-3.11, node-20 and claude-code kept current and snapshotted. We can build your own image into a template too.

Team controls

Accounts for everyone on your team, API tokens, per-user quotas, preview URLs for ports and an audit log of API calls.

Who it is for

Teams that ship code they did not write.

If your product runs model output, student code or customer scripts, you need a machine per run. We make that someone else's pager.

AI agent teams

A real shell and file system per agent session. Tool calls from Claude Managed Agents can run in OmniRun microVMs without you operating the worker.

Platform teams

An internal sandbox service for CI, evaluations and notebooks, with quotas per user and an audit log.

Agencies

Sandboxes inside client products, with a token and a quota per client, hosted in the EU.

How to start

A conversation, not a checkout.

Managed OmniRun is arranged per client or team. There is no public price list and no self-serve plan: we size limits with you.

Tell us what you run

Workload, expected volume, templates and where the data may live.

Talk to us

Agree limits and terms

Sandbox limits, quotas, custom templates and a price that fits the work.

Get your keys

Move in-house any time

Both run the same open-source release and API. Point your code at your own host and it keeps working.

Tell us what you are building.

Email hello@omnirun.io and we get back to you. The SDKs are the same for managed and self-hosted OmniRun.

Talk to us Read the docs

TypeScript$npm i @omnirun/sdk

Talk to us

Tell us what you want to run.

A person at A14A reads every message and replies by email, at hello@omnirun.io.

What runs inside. Agents, playgrounds, CI jobs, customer code.
How long it lives. Seconds, minutes or hours per sandbox.
How many at once. A rough peak is enough.
Network needs. Open internet, none, or a list of HTTPS hosts.

Prefer to run it yourself? Self-host OmniRun, the code is the same. Already have an account? It keeps working: open the dashboard.