Runtime as featured inForbesRead the article

Guardrails to control what every agent can do

We assume every agent could be compromised. So keys, network access, and approvals are enforced from outside its sandbox, and every step is on record.

Trusted by teams running mission-critical operations.

RainNixtlaMonoLineLeap

How every agent stays contained

Nothing inside the sandbox is trusted to enforce security. Each layer works on its own, so a poisoned prompt has to beat all of them.

1

A sandbox for every session

Every agent works on its own machine: a microVM or a gVisor-sandboxed container. Nothing is shared between sessions.

2

Real keys never enter the sandbox

The agent only ever holds a stub. A credential proxy outside the sandbox attaches the real key to approved requests, so a hijacked run has no key to steal.

3

Nowhere to send your data

Switch a sandbox to allowlist mode and agents reach only the hosts and IP ranges you approve. Everything else is blocked at the VM level.

4

A check on every tool call

Hooks run your own checks before and after each tool call, with a deterministic script or a model. Command deny lists always win.

5

A person approves what matters

Sensitive actions pause for a human sign-off from the dashboard or Slack before anything happens.

6

Every step on record

Every run is stored with its prompt, steps, tools, approvals, and cost. Searchable and exportable.

Diagrams are illustrative.

Specific claims you can check

Short and testable. Each links to the docs where we describe it.

Sandbox isolation

One machine per session, a microVM or a gVisor-sandboxed container. Nothing is shared between sessions.

Product architecture

Encryption

Tool connections and template secrets are stored encrypted. The API never returns the secret value.

docs.runtm.com/concepts/tools

Access control

Member, admin, and owner roles across the dashboard, CLI, and API. An API key can never exceed its owner’s role.

docs.runtm.com/concepts/guardrails

Audit logs

Prompts, tool calls, file changes, deploys, key usage, and every blocked command, with who, when, where, and cost. Exportable by API.

docs.runtm.com/concepts/guardrails

No training on your data

We never train on your data. Purge session data on the retention schedule you set.

Product policy

Secrets handling

Team secret values are write-only. Members see names, never values. Credentials go from the person to Runtime, never through a prompt or a repo.

docs.runtm.com/concepts/organizations

Guardrails and approvals

Deny, ask, or allow shell commands by pattern, enforced by a hook inside the VM. Production deploys and sensitive actions wait for a human.

docs.runtm.com/concepts/guardrails

Organization scoping

Every record is scoped to your organization. Bring your own AWS account to keep execution in your cloud.

Product architecture

Rules your agents can't talk their way around

Guardrails are configuration, not prompts. Set them once per organization, team, or agent.

RBAC for humans and agents

Roles and groups for your team. Agents get their own identity, scoped API keys, and a list of who can call them.

Budgets and spend limits

Cap cost per session and spend per user per day or month. Hard limits block, soft limits warn.

Command allow and deny lists

Allow, ask, or deny shell commands by pattern. Anything unmatched follows your default.

Deterministic scripts

Enforce rules with plain bash scripts, not prompts. The same input gets the same verdict every time.

Prompt injection

Built for agents that read untrusted content

Agents read emails, tickets, and logs that anyone can write. A poisoned instruction has nothing to grab and nowhere to go.

  • Content is evidence, not instructions

    Skills tell agents to treat emails, tickets, logs, and web pages as data to report on.

  • Nothing to steal, nowhere to send it

    Real keys stay in the proxy, and egress allowlists block every host you have not approved.

  • Least privilege, then approval

    Read-only credentials and deny lists limit what a poisoned prompt can reach. Real actions still need a human.

Built to stay up

A payout investigation shouldn’t stop because one cloud has a bad day. Sessions spread across providers and move when one degrades.

  • Multi-cloud sandbox redundancy

    Run on E2B, Google Kubernetes Engine, or AWS Lambda microVMs, and switch providers if one has an issue.

  • Multi-tenant by design

    Every record is scoped to your organization. Bring your own AWS account to keep execution in your cloud.

  • Your data stays yours

    We never train on your data. Purge session data on the retention schedule you set.

The same agents can defend you

A squad of agents watches your activity and infrastructure logs, investigates threats, and recommends a response for your team to approve. We run this on our own platform.

Explore cybersecurity agents
AICPA SOC 2

Request the SOC 2 Type II report

Our SOC 2 Type II report is available. Policies and controls are in the trust center.

Request the reportMonitored by

Data handling questions

Do you train models on our data?+

No. We never train on your data.

Who can see a secret after it is saved?+

Nobody through the API. Team secret and connection values are write-only: members see names and admins manage values, but values are never returned. Agents hold a stub, and a credential proxy outside the sandbox attaches the real key to approved requests.

Can an agent reach the open internet?+

By default a sandbox has outbound access to the public internet. Turn on allowlist mode and only the hosts and IP ranges you list are reachable, enforced at the VM level.

How long do you keep session data?+

You set the retention schedule, and session data can be purged on it. Paused sessions expire when their retention period lapses.

Can our own infrastructure run it?+

Yes. Use your own AWS account for execution, or self-host the control plane in your VPC and bring your own sandbox layer.

Self-host Runtime

For organizations that cannot send code or execution outside their own infrastructure. Run Runtime inside your VPC, on your compute.

Book a demo
  • Your cloud, your rules

    Run the entire Runtime control plane on your own infrastructure. Nothing leaves your network.

  • Bring your own sandbox

    Already run Firecracker, gVisor, or a custom execution layer? Runtime can use it instead of ours.