What R3 Daily Autopilot does, what counts as a billable task, what it costs, and the cap that cannot be exceeded.
Last updated 2026-09-24
R3 Daily Autopilot is a scheduled service that builds your actual plan for you, ten times a week, from the R3 material on your account and the situation you are actually in.
It is not a reminder, it is not a motivational message, and it is not a newsletter. Each run reads your configuration and your most recent check-in, decides whether today is a five, ten or twenty minute day, selects meals from recipes you own that fit your allergies, your household and what is in your cupboard, sets a walking target for the phase you are in, and sends you the result. That is work, and it is the reason it is billed per run rather than bundled into a flat subscription fee.
It is not one of the R3 Continuity gifts, and it is not R3 Coach AI. It is its own capability with its own pricing and its own cap.
A normal active week is ten scheduled runs.
Runs are scheduled server-side and are not affected by whether you are signed in, which browser you use, or whether your phone is switched on. Closing the tab changes nothing.
A task is billable when it completes and produces a real result: a plan containing meals and a session, delivered to you.
These are not billed:
Internally, a run reserves its price against your weekly allowance before it does the work, and releases that reservation if the work produced nothing useful. That ordering is what makes the cap below reliable.
$1.75 per completed task.
A normal week of 10 tasks is therefore $17.50.
Autopilot cannot bill you more than $25 in any weekly billing period, whatever happens. Once that is reached, no further billable tasks run until the next period begins. There is no override, no exception, and no setting that raises it.
This is enforced on our servers as a single atomic operation, so simultaneous runs, retries and out-of-order events cannot between them slip past it.
Billing weeks run Monday to Sunday, UTC. R3 Continuity and R3 Coach AI charges do not count toward this $25, and this cap does not limit them. They are separate lines with separate rules. R3 Continuity also happens to be $25 a week, but it is a membership price, not part of this cap.
Setup is short on purpose. These are the fields the planner genuinely cannot work without:
These are optional, and an empty one is never treated as a signal:
Allergies and foods you must avoid are applied as a filter in code, before the planner sees the list of candidate meals. They are not a suggestion the planner is asked to respect; they remove options from the menu entirely.
Your configuration is stored against your account and is editable at any time from your dashboard. It is never sent to any advertising platform. See our privacy notice for the full list of what is and is not shared.
In order, every time:
Steps three and seven are the pair that make the cap real. Reserving before working, rather than billing after working, is what stops a burst of simultaneous runs from each discovering room that only one of them actually has.
Runs are scheduled against your local time, using the offset recorded when you enabled Autopilot. Morning runs are built for early in your day; the three evening reviews land in your evening.
If infrastructure is briefly unavailable, a missed run is picked up on the next sweep where it still makes sense to send it, and is looked back on for a short window only. A morning plan that would arrive two days late is dropped rather than sent, because a plan for a day that has already happened is not a plan.
Every execution is recorded with its scheduled time, its actual start and finish, its status, its type, whether it produced a meaningful result, the price applied, its billing state, any retries, and whether the result email was sent. That record is what your dashboard displays and what any billing question is answered from.
Completed tasks accumulate through the week and are settled once, at the start of the following week, as a single charge described as "R3 Daily Autopilot, week of date", with the number of tasks included in the description. Settlement is idempotent: a retry cannot charge the same week twice.
Your dashboard shows, at all times: the current status, the configuration in force, the next scheduled run, how many tasks have completed this week, what they have cost, and how much of the $25 remains.
Each completed task sends one short email carrying that run's plan. Roughly ten a week on an active account, because there are roughly ten plans.
We do not send an email when you enable Autopilot, when you pause it, when you disable it, or when you change your settings. We do not send announcements, tips, nudges or offers through it. If an email arrives from Autopilot, it has a plan in it.
Unsubscribing from our emails also pauses Autopilot, because the result email is the product. You can switch it back on from your dashboard.
Autopilot requires a reusable payment method: a card, or a PayPal wallet authorised for future payments.
If you paid with Apple Pay or Google Pay only, Autopilot is not available. Both wallets provide a single-use device number, so there is nothing on file that can be charged again. On those accounts Autopilot is not offered, not shown, not scheduled, and blocked at the route. Nothing you bought is affected.
You can add a card deliberately from Billing at any time.
Pause stops runs immediately and keeps your configuration. Disable stops runs and switches it off; nothing re-enables it automatically, and if you disable it you will not be asked again by the system.
If a weekly settlement fails, the outstanding amount is shown in your dashboard and further billable tasks are held until it clears. Everything you have bought stays available regardless.
Autopilot builds an education and movement plan. It is not medical care.
If your check-in mentions anything that belongs with a clinician, Autopilot does not build a more demanding plan. It drops to the shortest, most conservative option, surfaces the relevant guidance, and names the professional to speak to. It will not interpret what you reported as a condition, because that is diagnosis and it is not ours to do.
If you have not confirmed that you are cleared for normal activity, Autopilot plans walking and breathing work only, regardless of what else your configuration says.
Plans are generated automatically and select from the R3 material on your account. They are generated from what you told the system, so a configuration that is out of date produces a plan that is out of date; keeping it current is the one thing the service needs from you.
Autopilot can be wrong. It is a planning aid, not an authority, and the R3 product itself is the reference whenever the two differ. Use your own judgement, and skip anything that does not feel right on the day.
The service depends on infrastructure and on a third-party AI provider, and can be delayed or briefly unavailable. A run that does not happen is not billed. Missed runs are picked up on the following sweep where it still makes sense to do so.
Keep your configuration accurate, especially your allergies, anything you must avoid, whether you are feeding, and whether you have been cleared for activity. Read what arrives before you act on it. Stop anything that hurts. Take anything about your body or your baby to a qualified professional rather than to an automated plan.
Anything about a charge, a run that did not arrive, or a plan that looks wrong: support@r3-method.com.