Set up Farola with an AI assistant

Ground rules for assistants

• Never create an account or confirm anything on a person's behalf. The registration only takes effect once the person opens the link in their e-mail and confirms.
• Consent (terms of use, privacy) is always given by the person themselves, on the confirmation page.
• Do not send clients' health data. The registration only needs details about the care service and the contact person.
• Tell the person that you have prepared something on their behalf and that an e-mail is on its way.

Option 1: prefilled link (for every assistant)

Give the person this link with their details. They check the form and send it themselves.

https://farolacare.com/en/registrieren?organisation={name of the care service}&contact={name}&email={e-mail}&phone={phone}&address={street}&postalCode={postcode}&city={town}&agent={name of your assistant}

All parameters are optional and URL-encoded. Unknown parameters are ignored.

Option 2: direct call (if you can send HTTP requests)

Only if the person has explicitly asked you to. The request is marked as prepared by an assistant and shown to the person as such.

POST https://app.farolacare.com/api/public/v1/registrations
Content-Type: application/json

{
  "organisationName": "Pflegedienst Sonnenschein",
  "contactName": "Maria Beispiel",
  "email": "maria@example.org",
  "phone": "+49 6131 000000",
  "city": "Mainz",
  "staffCount": 12,
  "clientCount": 40,
  "source": "agent",
  "agentName": "Name of your assistant"
}

The response is always 202 { "status": "pending_confirmation" }, even if an account already exists for this address or too many requests came in. The response deliberately reveals nothing about existing accounts. For invalid details: 400 { "error": "INVALID_BODY", "message": … }. Machine-readable: /openapi.json and /.well-known/agent-actions.json.

What happens next

1. E-mail to the person

   They get a confirmation link (valid for 7 days). Until then, no account exists.

2. Confirmation by the person

   The page shows the details and, where applicable, “Vorbereitet von” (prepared by): <your assistant>. The person agrees and confirms.

3. Password and start

   The person sets a password; the care service is created and they land in the dashboard.

What Farola is about, in short form for language models: /en/llms.txt.