How automation billing works
Unlimited moderation, pay for engagement. Your plan meters automations — and only the ones that reach somebody on your behalf: a public reply, a DM, a like. Hiding, deleting, approving and promoting are never metered, at any volume. Neither is anything you do by hand.
Ingestion is unlimited. Comments and DMs arriving are not metered at all — there is no allowance on how many the product pulls in, reads, and analyses for you. The only thing that moves your meter is the product acting outward.
When you reach your plan’s included automations, ingestion still does not stop. New comments and DMs keep arriving, your Inbox stays live, and manual moderation always works. What pauses is automated processing — your Workflows stop running — unless the billing owner has switched on extra usage, in which case automation keeps running and each metered automation past the allowance is billed at your plan’s per-automation rate. Nothing is ever billed without that explicit opt-in.
What counts toward your usage
Section titled “What counts toward your usage”The metered unit is an automation the product performed on your behalf. Every workflow action falls on one side of this line, and the line does not move with volume:
| Action | What it does | Counts toward your allowance? |
|---|---|---|
| Flag | Hides the comment on the platform — Flag is the hide | Never |
| Delete | Removes the comment, where the platform supports it | Never |
| Allow | Approves the comment, and un-hides it if it was hidden | Never |
| Promote | Marks the comment internally; no platform change | Never |
| Reply | Posts a public reply on the post | Metered |
| Private reply (DM) | Sends a direct message to the comment’s author | Metered |
| Send message | Sends a direct message to a recipient | Metered |
| Like | Likes the comment on the platform | Metered |
| Reply with AI | Runs the AI model to draft a reply | Metered |
The four on top are defensive — moderation, the work of protecting your page. The five below are the product doing something for you: four of them speak to a follower in your name, and the fifth runs the AI model on your behalf. You pay for that kind and never for the first.
Three more things are never metered, for reasons that have nothing to do with which action ran:
- Anything a human did. Approving, rejecting, hiding, liking or replying by hand from the Inbox — including DM replies you type in the composer — is your work, not the product’s. It is free, and it keeps working through every pause.
- Anything in observe-only mode. A Workflow running in observe-only (shadow) mode simulates its actions and sends nothing — and never calls the AI model — so it bills zero at any volume. Testing a Workflow is free by construction.
- A send the platform refuses, or a generation that fails. Metering happens when the platform accepts the action, not when the product attempts it. A reply blocked by an expired token bills nothing, and neither does an AI draft the provider failed to produce.
An AI draft is an automation
Section titled “An AI draft is an automation”Reply with AI was free until recently, on the grounds that it drafts text and sends nothing. That was the wrong way to look at it: the AI model ran, and running it is the work you are paying for. So it costs one automation per generation, out of the same monthly allowance a reply or a like comes out of — Base 1,000, Pro 5,000, Premium 25,000. There is no separate AI quota and no weekly limit any more.
Four things follow from “one generation, one automation”:
- Wherever you asked from. A workflow’s Reply with AI branch and the Generate AI draft button in the composer cost exactly the same, because they are the same work.
- Whether or not you send it. A draft you read and discard costs one. A held draft waiting for your review costs one the moment it is generated.
- Never two. If you do send the draft, the send is metered as the reply it is — the generation is not charged again.
- A failure costs nothing. If the provider is down or times out, no automation is spent.
Two details on how metered automations are counted:
- One action counts once, even when it produces both a platform send and a push notification, and even if several Workflows or branches evaluated the same comment.
- Counted by when the automation ran, not when the comment was posted. A reply to last month’s comment counts against this month’s window.
Your plan’s included automations
Section titled “Your plan’s included automations”Each plan includes a fixed number of metered automations per month:
| Plan | Automations per month | Price |
|---|---|---|
| Trial | 700 | Free, 10 days (a preview of Base, with Pro-level features) |
| Base | 1,000 | $29/mo · $278/yr |
| Pro | 5,000 | $79/mo · $758/yr |
| Premium | 25,000 | $199/mo · $1,910/yr |
That number is the whole allowance. Ingested comments and DMs, moderation actions, Workflows, connected accounts, and client workspaces are all unlimited on every paid plan — there is no cap on how much the product reads for you, how much of it you hide, how many Workflows you build, or how many Facebook Pages and Instagram accounts you connect.
Seats are capped, and they are the exception. Every plan includes a fixed number of team seats including the owner: Base 1, Pro 5, Premium 15. Base is a single-owner plan with no team features (it also has no webhooks, smart analysis, intent rules or knowledge base — connecting your website so replies can answer from your own pages is included from Pro). AI drafting is not one of those differences — every paid plan can draft, and a draft costs one automation on all of them. Pending invitations hold a seat until they are accepted or expire, so a Pro team with three members and two outstanding invitations is full. Seats are not purchasable individually — a sixth seat means Premium.
What happens at your allowance
Section titled “What happens at your allowance”State this plainly, because this is the part people worry about: at your allowance, we never stop listening, and we never bill without permission. What happens next depends on one switch — whether the billing owner has enabled extra usage.
Without extra usage (the default):
- Automated processing pauses. Your Workflows stop running. Nothing is billed — there is no automatic overage.
- The free actions pause too, and that is the real cost of a pause. The pause stops automation, not billable automation, so automatic hiding and deleting stop with everything else — not because they cost anything, but because your Workflows are not running. Hiding by hand still works.
- Ingestion continues. New comments and DMs keep being pulled in and land in your Inbox as usual. Nothing is dropped or missed.
- Manual moderation always works. Replying, hiding, and liking from the Inbox keep working through a pause, and stay free.
- Resuming catches up. When capacity returns — your window resets, you upgrade, or extra usage is enabled — automation resumes and catches up the last 24 hours of queued comments and DMs.
With extra usage enabled:
- Automation keeps running. Past your included automations, each metered automation is billed at your plan’s per-automation rate (see below); the defensive ones stay free past the allowance exactly as they were inside it. If you set a spend cap, automation pauses when accrual reaches it.
The in-app banner during an allowance pause reads: “Automation is paused at your plan allowance”, with a live count of what is queued for your rules and the reminder that replying, hiding and liking keep working, and stay free.
Extra usage: opt-in, per-automation billing
Section titled “Extra usage: opt-in, per-automation billing”Extra usage is off by default and can only be switched on by the billing owner — the flip of the toggle is the consent, and the card records who enabled it and when. No deploy, checkout, webhook, or background process can turn it on for you. (It is not available during the trial; the trial simply pauses automation at its 700-automation allowance.)
The threshold and the charge are now the same unit, which is what makes this simple to reason about: your included automations decide when extra usage begins, and past that point each further metered automation costs the rate below. You are never billed for a comment arriving, for reading it, for analysing it, or for hiding it.
The rate is a fixed price per automation, set per plan. These three figures are the published numbers and the ones worth remembering — they are also exactly what your invoice is raised at, to the cent:
| Plan | Included automations / month | Rate per extra automation |
|---|---|---|
| Base | 1,000 | $0.05 |
| Pro | 5,000 | $0.03 |
| Premium | 25,000 | $0.01 |
Extra usage is deliberately priced above the plan itself: sustained overage means upgrading is the cheaper path, and the app tells you when that crossover happens.
- Spend cap (optional). The billing owner can set a dollar cap on extra-usage accrual: we alert at 80% of the cap and pause automation at 100%. With no cap set, automation never pauses for spend. Lowering a cap below what has already accrued pauses automation immediately — the app asks you to confirm before saving that.
- Disabling is not a refund. Turning extra usage off stops new accrual; anything already accrued stays on your next invoice.
- Plan changes are prospective. Changing plans mid-window changes the rate for automations from that point on — accrued charges are never re-priced.
A worked example month
Section titled “A worked example month”You are on Base (1,000 included automations), with extra usage off. Your Workflows hide spam, reply to product questions, and like praise.
-
Days 1–12 — 9,400 comments and DMs ingested. All of it free: ingestion is unlimited. Your Workflows hide 3,100 comments and delete 40 — zero metered, because moderation never is. They also post 610 public replies and 150 likes: 760 metered. You are at 760 / 1,000 (76%).
-
Day 13 — a spam wave. 2,000 junk comments land overnight and your Flag branch hides every one of them. Your meter reads 760 / 1,000. It did not move. The wave cost $0.00, and it would have cost $0.00 on a spend cap of any size.
-
Day 15 — you cross 800 (80%). A dismissable notice appears in-app and we email the billing owner, with your days remaining in the window and what your options are. You keep moderating normally.
-
Day 22 — you cross 900 (90%). The notice becomes persistent (it stays until the window resets or you act), because the 100% consequence is real: automation will pause — including the free defensive actions.
-
Day 26 — the 1,000th metered automation runs. Automation pauses. Ingestion continues — new comments and DMs keep arriving in your Inbox, the queued-for-your-rules count ticks up, and replying, hiding and liking by hand keep working, free. Automatic hiding stops until the pause lifts.
-
Window resets (or you upgrade, or enable extra usage). Automation resumes and catches up the last 24 hours. Had extra usage been on, automation would never have paused — the first reply past 1,000 would have cost $0.05, and the spam wave would still have cost nothing.
At no point in this example are you charged anything beyond your $29 plan — and the single busiest, ugliest night of the month was the one that cost the least. Without the opt-in, the allowance is a pause on automation, never a surcharge; and even with the opt-in, the bill is per outbound automation, never per comment that arrives.
Automation can pause for three reasons
Section titled “Automation can pause for three reasons”When automated processing is paused, it is always one of exactly three states, each with its own way back:
| Pause reason | What happened | What resumes it |
|---|---|---|
| Allowance | Included automations used up, extra usage off | Window resets, you upgrade, or the billing owner enables extra usage |
| Spend cap | Extra-usage accrual reached your cap | Raise or remove the cap, upgrade, or the window resets |
| Past-due grace | A payment failed; extra usage behaves as not-enabled during the 7-day grace window | Update the payment method, or the window resets — automation resumes on the fresh allowance, while extra usage stays paused until the card is fixed |
In every one of these states, ingestion continues, the Inbox stays live, and manual moderation keeps working. Resuming catches up the last 24 hours.
Upgrading, downgrading, and payment problems
Section titled “Upgrading, downgrading, and payment problems”Plan changes go through Stripe Checkout.
- Upgrades take effect immediately. Once checkout completes, your new plan and its larger allowance apply right away. If automation was paused at your old allowance, it resumes and catches up the last 24 hours.
- Proration is handled by Stripe. We do not do proration math in-app — Stripe calculates any credit or charge for switching mid-period. Extra-usage charges already accrued are never re-priced by a plan change.
- Downgrading pauses seats — it never deletes anything. A downgrade is blocked while extra seats or plan-gated webhook endpoints are active. When a downgrade would take your team over its new seat count, the owner chooses which seats stay via the seat picker; the remaining non-owner memberships are paused, not removed — the member rows, their roles, your Team Workspaces and all history stay exactly as they were. A paused member sees an explanatory screen rather than a missing workspace, and re-upgrading restores everyone in one step with nothing to re-invite or rebuild. Webhook endpoints beyond the new plan’s entitlement get a 30-day grace period, then pause — they are never deleted.
- If a payment fails, service is not interrupted. You get a 7-day past-due grace window to update the payment method. Everything keeps working on your plan’s included allowance; extra usage simply behaves as not-enabled until the card is fixed.
- Monthly and Yearly are billed separately. Yearly is offered at roughly a 20% saving versus monthly.
Where to see your current usage
Section titled “Where to see your current usage”-
Open Account Settings.
-
Go to the Billing & Usage tab (this is the default tab).
-
Read the Extra usage card. Its meter has two segments: the part of the bar inside your allowance (which never charges — it is labeled $0.00) and, once you are past the allowance, a second segment for what is beyond it. When extra usage is enabled, a second meter shows the automations being billed and your accrual against the spend cap.
In-app notices step up at 80% of your allowance (a dismissable notice), 90% (persistent), and 100% (the pause itself, with its reason and the way back), and they appear for everyone approaching the allowance. Emails follow the same 80 / 90 / 100 rungs only while extra usage is not in force (off, or suspended during the past-due grace) — an opted-in payer gets the spend-ladder emails (cap alert, milestones) instead. Unlimited items (ingested comments and DMs, Workflows, connected accounts) show as Unlimited rather than a number.
The same panel shows your current plan, billing history, and a Change Plan control for upgrades and downgrades.
Related
Section titled “Related”- Extra usage and spend caps — the opt-in, the rates, and the cap in detail
- What happens when processing pauses — the three pause reasons, what keeps working, and the catch-up
- Which actions are metered — the defensive/outbound split, action by action
- Team seats and the seat picker — the seat ceiling and what a downgrade does to members
- Webhooks by plan, and the downgrade grace
- What we access and store from Meta
- What happens while an account is disconnected
- How Workflows work
- Actions: Flag, Allow, Reply, and Delete