ManyChat Pricing Explained: Why the Unit Matters More Than the Number

Most people looking up ManyChat pricing want one number. The number is the least useful part. ManyChat bills per contact, and on Instagram the definition of a contact is where the whole cost model lives — the per-unit rate is almost a rounding detail next to it.

This is not a complaint about ManyChat specifically. Per-contact pricing is now the default across the category, and several of the tools marketed as cheaper alternatives use exactly the same unit with a smaller multiplier. Understanding the unit is what lets you predict an invoice; understanding the rate is not.

What counts as a contact on Instagram?

A contact is created the moment a person becomes addressable by your automation. On Instagram, that means any of these:

• Someone comments on your post and your comment-to-DM rule answers them.
• Someone sends you a first Direct message and any automation responds.
• Someone mentions you in a story and your story-reply rule fires.
• Someone clicks into a flow from an ad or a link and lands in Direct.

Notice what is not on the list: buying something, replying twice, opting in, or staying interested. A contact is a person your system said one sentence to. Once created, they stay counted for the billing period regardless of what happened next.

This matters because of a rule Meta enforces and no tool can change: you may reply to someone for 24 hours after their last message, and you may send one private reply per comment within seven days. After those windows close, a contact is not a subscriber you can reach. They are a row on the invoice you are not permitted to write to. The full rule set is in the rules that govern Instagram DM automation.

How the invoice is actually built

Three mechanics combine, and they combine badly on a good month:

1. The count is cumulative, not concurrent. You are not billed for people currently in a conversation; you are billed for everyone who entered one during the period. Volume is a ratchet.

2. Tiers are stepped. Crossing a tier boundary by a handful of contacts moves you to the next band for the whole month. The cost curve is a staircase, and one strong post can carry you up two steps.

3. A viral post is the worst case, not the best. A Reel that outperforms by 10× produces roughly 10× the contacts. Revenue from that Reel might be up 10× too, or it might be up not at all — reach and sales are only loosely connected on Instagram. The bill does not wait to find out.

The concrete version: a comment-to-DM rule on a Reel that gets 600 comments creates roughly 600 contacts in the hour the Reel peaks. If your normal month is 400 contacts, one good afternoon more than doubled your bill and you will not know for weeks whether it paid for itself.

Is the ManyChat free plan usable in 2026?

As of March 2026 the free allowance is 25 contacts. That is not a starter tier; it is a demo. Twenty-five contacts is enough to test a keyword with a colleague and a second account, and that is the honest description of it. The old plan — start free, validate, upgrade when it clearly works — no longer exists, because 25 contacts does not produce enough evidence to validate anything.

If you are searching for ManyChat pricing because the free plan stopped working, the thing to compare is not another free tier with a slightly larger number. It is whether the next tool bills on the same unit, because if it does you have bought yourself a few months, not a different situation.

When does per-contact pricing actually make sense?

It is worth saying plainly: sometimes it does. Per-contact works when the contact is genuinely an asset — when your channel lets you re-engage on your own schedule, when each new contact has a knowable expected value, and when volume grows roughly in step with revenue.

That describes email. It describes SMS. It describes a Facebook Messenger list under some conditions. It does not describe Instagram, where re-engagement is what Meta specifically prohibits, and where a contact’s addressable life is measured in hours.

What to compare instead of price per contact

If you are mid-way through a comparison, these four questions predict the next twelve months better than any rate card:

1. What is the unit? Per contact, per account, per seat, per plan. Ask what a spike does to it.
2. What happens at the ceiling? Does the automation stop answering, does it keep running and overage appears, or are you upgraded automatically? All three exist, and only one of them is discoverable from a pricing page.
3. What do you lose on downgrade or cancellation? Specifically: conversation history and the flows themselves.
4. What is included versus metered? AI answers in particular are frequently a separate meter, and the interesting number is what an AI reply costs at your volume, not whether AI is “included”.

How Chatonaut prices this differently

Chatonaut is not billed per contact. A plan covers the account, and a post that does unusually well does not change what the month costs — which is the whole point, because you cannot predict which post that will be. What each tier includes is on the pricing page, and the free plan is sized to run a real automation, not to demonstrate one.

The rest of the comparison — inbox, AI agent, handover — is in the tool-by-tool comparison, and if you are specifically weighing a move away from ManyChat, the 2026 breakdown covers what changed and what breaks when you switch.

If you have not set up a comment-to-DM rule yet, start with the comment-to-DM guide — it is the automation that produces most of the contacts you will be billed for anywhere, so it is worth understanding before you pick a vendor.