ManyChat or the WhatsApp Business app: which one do I need?

If you are one person, or a business of two, answering from a phone and your day is long enough to reply to everybody, the WhatsApp Business app is the answer and you need nothing else. It is what Meta says about its own product: the app is for small businesses that handle things personally, and the API is for medium and large businesses that need programmatic access. When the day stops being long enough, that sentence becomes the diagnosis.

If your traffic comes from Instagram and ads, and what you want is to automate replies, capture contacts and send them sequences, ManyChat does that well and it is where it was born. The fine print is twofold: WhatsApp is not on the lower plans — it appears from Pro at $39 USD per month — and the price rises with your active contact base rather than with your features.

If you answer from several phones, or two people fight over the same number, neither is designed for that. The app supports one primary phone plus up to four linked devices; ManyChat solves automation, not the running of a sales team with history, stages and owners.

And if your conversation has to search a catalog, quote with terms, check a calendar or take payment, the comparison changes category entirely: that is no longer automating messages, it is doing work inside the conversation.

The WhatsApp Business app appears on no rate card: you download it and that is that. Its cost is the time of whoever answers, and that cost becomes visible in the month somebody spends three hours a day on the phone repeating themselves. It is worth calculating even though it shows up on no receipt.

ManyChat does publish a rate card, and its ladder is built on active contacts. The free plan carries twenty-five contacts, four automations, two channels and ManyChat branding in the conversations, and it does not include WhatsApp. Essential costs $17 USD per month — or $14 USD paid annually — with two hundred and fifty contacts, and it does not include WhatsApp either. Pro costs $39 USD per month, or $29 USD annually, with two thousand five hundred contacts, and it is the first plan where WhatsApp exists. Business costs $99 USD per month ($69 USD annually) with seven thousand five hundred contacts, and Advanced $199 USD per month ($139 USD annually) with twenty-five thousand.

The detail that decides the real budget is the extra active contact, and its price moves by plan and billing cycle: $0.10 USD on Essential and $0.05 USD on Pro with monthly billing, with lower rates on the higher plans. Five cents sounds like nothing until an ad campaign brings five thousand new people in a month; on Pro, that is two hundred and fifty dollars on top of the subscription. ManyChat's price is decided not by what you do but by how many people write to you.

On currency, ManyChat bills in dollars only, with no peso price and no Mexican tax-invoice commitment, and Meta's per-delivered-template fees sit outside. Budgeting this in Mexico means adding three things: the dollar subscription, the extra contacts and Meta's fees.

The WhatsApp Business app has no agent: a person answers. Its tools help you type faster, but nobody replies for you while you sleep. If what you want is repeated questions answering themselves, the app is not where that discussion happens.

ManyChat includes artificial intelligence from its Pro plan, the same one where WhatsApp appears. What it does not publish is the allowance: it does not say how many replies are included, what the overage costs or what happens the day it runs out. That does not make it bad, it makes it hard to budget, and it is exactly the question to send in writing before signing.

The other cost, on neither page, is maintaining the automations. An automated flow works as long as the world holds still: the moment your prices, hours or stock change, somebody has to go in and update it. That recurring work is the expensive part of automating with flows, and it appears in no price comparison.

The practical boundary is this: if your questions are five and always the same, an automated flow handles them and you do not need an agent. If the answer depends on data that moves — what is available, at what price, at what hours, on what terms — the flow breaks and the conversation ends in «let me connect you with an advisor», which is exactly where you were before paying.

The official app rests on a phone. Meta documents one primary phone plus up to four linked devices, and that those devices are logged out if the phone goes fourteen days without connecting. That is the whole architecture, and also its ceiling: the number lives in a device somebody has to keep charged.

For sending to many people, the app carries native broadcast lists with two conditions that surprise nearly everybody: the recipient must have you saved as a contact to receive the message, and each list has a contact ceiling. We do not publish the exact number here because we could not re-verify it on Meta's own page — its help center blocks automated readers — and a second-hand figure on a page like this is worth nothing. The rule itself is documented, and it is what breaks most broadcasting plans.

That route is also narrowing: Meta is already converting broadcasting into a paid business send in some countries. This comes from a Meta help-center answer archived in July 2025, so check the official page for your country on the day you need it rather than assuming it is still free.

ManyChat takes the other road: it connects to the official API from its Pro plan and its broadcasts run on templates, with Meta's per-delivered-template fees outside the subscription. That removes the requirement that the customer has you saved and removes the per-list ceiling, but it introduces a per-template-message cost the app does not have. That is the real trade in this comparison, and neither page states it in a single sentence.

Eva is our product and it enters here as a third option, with a public price: from $191 USD per month on an annual plan ($229 USD month to month) on the Esencial plan, with 5 users and 5,000 credits included, and artificial intelligence already inside the price. An extra user costs $17 USD per month and a 1,000-credit top-up $49 USD. We do not bill per contact: twenty thousand people can write to you and the meter stays the same.

Where Eva beats both is in what happens inside the conversation. The app depends on a person being there; ManyChat automates messages with flows somebody has to maintain. Eva puts an agent that quotes from your real catalog with stock and terms, books against true availability, takes payment inside the conversation, invoices in your customer's name and follows up for months, and all of it lands in the built-in CRM with nobody typing it in. And it handles WhatsApp, Instagram, Messenger, calls, email and your website, plus Instagram comments and TikTok ad comments.

The administrative side changes too: a peso price plus VAT inside Mexico with Mexican tax invoicing, instead of a dollar subscription with per-contact charges discovered at the end of the month.

Now the honest part. If you answer from a phone yourself and your day is long enough, the official app is free and sufficient, and paying for any platform on top would be throwing money away: stay where you are until the day runs short. If your job is capturing contacts on Instagram and sending them sequences, ManyChat is built for that and costs less than Eva on the lower plans. And if you need to sign up alone today, ManyChat has a free plan and Eva has neither self-serve signup nor a free trial: guided setup is prepared from your site and your catalog, you approve it at no cost, and assisted setup is bought separately from $579 USD.

First, measure the day. Count how many minutes somebody on your team spends answering messages today, and how many messages arrive when nobody is there. If the answer is «almost none», the official app wins and this comparison is over, free of charge.

Second, count contacts, not features. If you are going to compare ManyChat, take how many distinct people wrote to you last month and place it on its ladder: two thousand five hundred, seven thousand five hundred, twenty-five thousand. Everything above that is billed at five cents per contact, and that is where the price stops resembling the front page.

Third, check whether your broadcasting plan survives the rules. If the idea was to send promotions from the app, verify two things first: that your customers have you saved as a contact, and what Meta is charging for broadcasting in your country today. If your plan does not survive those two questions, you already know you need the official API.

Fourth, read fifty real conversations from last week and mark how many are settled with a fixed message and how many need searching, quoting, booking or charging. If nearly all are fixed, automate with flows and do not pay for an agent. If half need real work, you know which category you are buying.

Fifth, ask in writing what the pages do not say: how many artificial-intelligence replies the plan includes, what happens when they run out, how an active contact is counted, and which Meta fees stack on top. Four written answers order the decision better than any table.

• ManyChat's figures were verified on 14 August 2026 against the official archived July 2026 copy of its pricing page, because the live site rejects automated readers. That is a second-best source compared with the live page, which is why it is declared here rather than hidden.
• The WhatsApp Business app publishes no rate card; its positioning against the API is quoted from the official WhatsApp Business page, and the limit of one phone plus four linked devices from the WhatsApp help center.
• The per-list contact ceiling for native broadcast lists is not published here: we could not re-verify it on Meta's own page, and we would rather describe the rule without the figure than repeat a third-hand number.
• The conversion of broadcasting into paid business sends is documented in a Meta help-center answer archived in July 2025 and applies only in some countries; check the official page for your country before planning a campaign.
• Every figure is quoted in the currency its vendor publishes, with no conversion to another currency.
• Eva's figures come from our public rate card and are marked as ours wherever they appear.