Capabilities

Every run leaves a record.

Health data needs an answer for the auditor, not just the buyer. Each organisation runs in its own store, every key is scoped and revocable, and every run keeps its input, steps, tool calls, logs and output.

Your own store.

A brand’s loops, runs, users, scans and keys live in that organisation’s store, apart from every other customer.

Keys that stop.

An API token reaches only the loops in its organisation and can be revoked without touching the loops.

Secrets stay secret.

Credentials are encrypted at rest, shown once, and visible only to owners and admins.

The record is complete.

Input, output, status, every step, every tool call, every log line and every error, kept for the run.
How it works

The mechanism, in order.

  1. A person or system authenticates. A dashboard session or a bearer token scoped to all loops or one loop, with run access only.
  2. The organisation is the boundary. The org id in the trigger URL and the token have to match. A token from one org cannot start another org’s loop.
  3. The run is recorded. The platform writes the trigger, status, input, output, every step and its return, tool calls and console lines against the run id.
  4. A person reviews it. The Runs tab shows each run; the detail page shows its timeline. The Logs tab shows the stream, filterable by run.
  5. Data terms sit around it. POPIA and GDPR processing, a DPA, subprocessors listed in it, and the data location agreed on request.
run-record.jsonJSON
{
  "id": "run_91",
  "status": "succeeded",
  "triggerType": "api_token",
  "input": { "patientId": "pat_18", "phoneNumber": "+15551234567" },
  "output": { "patientId": "pat_18", "scanId": "scan_91" },
  "steps": [
    { "name": "Generate a health scan", "status": "succeeded", "return": { "scanId": "scan_91" } },
    { "name": "Send the report", "status": "succeeded", "return": { "messageId": "msg_44" } }
  ]
}
The specifics

What you get, and where it stops.

Isolation. Loops run in a sandboxed runtime and reach only the tools and hosts switched on for them. Each organisation has its own data plane for loops, versions, runs, users and keys. The global database holds identity and the pointer to that data plane.
Identity and roles. Organisation members sign in through the product. API tokens are organisation-scoped and revocable. Enterprise plans add single sign-on. Secrets are visible only to owners and admins.
Privacy. POPIA and GDPR processing, a data processing agreement, subprocessors listed in it, and data location on request. Customer data is never used to train models. The model sees only what the turn needs. Unsaved guest records are deleted after seven days.
Audit record. Every run stores its trigger type, status, input, output and timestamps; every step stores its name, status, timing, return value and error; the event log stores every console line by level; the tool trace stores each tool call made inside a step.
Deployment. The runtime runs on Cloudflare Workers, Workflows, Durable Objects and R2. There is no self-hosted or on-premises edition. A single named region is not offered; the data location is agreed in the DPA and available on request.
Where it stops. The security page does not publish a certification list or a subprocessor page outside the DPA. Ask for the current report and the subprocessor list in procurement; they are not claims on this page.
FAQ

Questions

Where is our data stored?

In the organisation’s own store on Cloudflare’s platform. The specific location is agreed in the data processing agreement and available on request; Loops does not offer a region picker.

Can we self-host?

No. Health Loops is a managed service. Enterprise plans include your own data plane, volume rates and single sign-on, not an on-premises install.

Do you have a DPA and a subprocessor list?

Yes. Every business can sign a data processing agreement, and the subprocessors are listed in it. Ask for it through the security page.

Who can see our secrets?

Owners and admins. Secret values are encrypted at rest, never shown again after saving, and read by loop code as environment variables.

How long do you keep a run?

Runs are kept and queryable. The run record includes the input, output, steps, tool calls and logs for the run.

Ollie’s agents give wellness guidance. They don’t diagnose or prescribe, and they hand off to a human expert when it matters.