The WhatsApp Business API has no single price. Meta charges per message or per template, with rates that differ by category (marketing, utility, authentication, service) and by country, published on Meta's own rate cards and revised over time. Your real total is Meta's fees plus your provider's margin plus your software subscription.
Why can nobody quote you one number?
Because the model is multiplicative, not flat. Four factors set your bill:
-
Message category. Marketing templates cost the most, utility (transactional) templates less, authentication messages are priced for OTP volume, and service conversations (replying to inbound customers) are the cheapest tier.
-
Country. Meta publishes a rate card per country, and the same marketing template costs meaningfully different amounts in Saudi Arabia, Egypt, and the US. The rates change, so any specific number in a third-party blog post, including this one if we quoted any, would eventually be wrong. Always check Meta's current published rates for your market.
-
Volume and mix. A store sending 10,000 shipping updates has a very different bill from one blasting 10,000 promotions.
-
Who sits between you and Meta. Access through the Cloud API is free from Meta; business solution providers (BSPs) add their own margins, platform fees, or markups per message.
How does the pricing model actually work?
Two mechanisms drive everything:
The 24-hour customer service window. When a customer messages you, a 24-hour window opens in which you can reply free-form, without templates. Each new customer message resets it. Outside the window, you can only initiate with pre-approved template messages, and those carry the per-template charges. Practically: inbound-driven customer service is the cheap end of the model, business-initiated outreach is the expensive end.
Category-based template pricing. Every template is classified:
| Category | Typical use | Relative cost |
|---|---|---|
| Marketing | Promotions, cart recovery, win-backs | Highest |
| Utility | Order confirmations, shipping updates | Lower |
| Authentication | One-time passcodes | Priced for scale |
| Service | Replies within the window | Lowest tier |
Meta has continued to evolve this model (the industry direction has been toward per-message pricing rather than per-conversation), which is one more reason to read the current official rate card rather than cached summaries.
What does a realistic budget look like?
Build it as three lines:
-
Meta fees. Estimate your monthly template volume by category. A typical Shopify store's mix: utility templates for every order (confirmation, shipped, out for delivery), marketing templates for cart recovery and campaigns, and a large mass of service conversations that cost little. Multiply by Meta's current rates for your customers' countries.
-
Provider margin. Ask any BSP one direct question: "Do you pass Meta's rates through at cost, and what do you charge on top?" Answers vary from transparent pass-through plus subscription to opaque per-message markups. Get it in writing.
-
Software. The platform that manages templates, conversations, and Shopify integration. This is usually the largest controllable line, and the one where value differs most: an inbox tool saves organizing time, an AI agent saves answering headcount.
Cost-control levers worth knowing: design flows so customers message you first where possible (opens the free-form window), prefer utility templates over marketing where the content is genuinely transactional, keep quality ratings high (blocked and reported messages waste spend), and do not send a third cart reminder where the second already failed.
Where do the savings actually come from?
Here is the part cost calculators miss: for most stores, message fees are noise next to labor. The expensive thing was never the template; it is the human hours answering the replies, or the sales lost when nobody does.
Data points from ReplAi's world: WISMO alone commonly runs 25 to 50 percent of ticket volume (industry reporting), and ReplAi's integration with Bosta in Egypt shows stores doing 1,000+ orders a month saving 4 to 8 hours of team time per day with 70 to 80 percent of messages handled without humans. On the revenue side, ReplAi's pilot cohort lifted DM-driven sales 150 percent at the median. Against numbers like that, the per-template fee is a rounding error; judge the whole stack by conversations resolved and orders closed, not cents per send.
Country-specific setup notes live in our guides for Saudi Arabia and the UAE, and the wider strategy in the complete WhatsApp for Shopify guide.
What does the software layer cost?
For orientation, ReplAi's pricing is public and volume-based: Free (100 replies a month), Starter at 24.99 dollars (500 replies), Grow at 79.99 (2,500), and Pro at 299.95 (10,000), with Meta's fees separate as always. ReplAi is an AI agent for Shopify rather than a bare API pipe: it answers on WhatsApp, Instagram DMs, Messenger, and WebChat in under 14 seconds on average, 24/7, automated 80 percent of cohort messages in Q1 2026, speaks Egyptian, Gulf, and Levantine Arabic plus Franco-Arabic and English, and executes Shopify actions in chat, from catalogue search and order tracking to placing COD orders. Setup is about 10 minutes. Compare tiers at pricing or book a demo to size it against your volume.
Frequently asked questions
Is the WhatsApp Business API itself free?
Access via Meta's Cloud API is free: no license fee for the pipe. You pay per message or per template you send by category and country, at Meta's published rates, plus whatever a BSP charges if one sits in the middle, plus your software platform's subscription.
Which messages are the cheapest to operate?
Service conversations, replying to customers inside the 24-hour window, sit at the bottom of the cost model. This is why support-led and inbound-led WhatsApp strategies cost far less per conversation than broadcast-led ones, and why answering fast (which keeps threads inside open windows) is also the frugal choice.
Where do I find the actual current rates?
On Meta's official WhatsApp Business Platform pricing pages, which publish rate cards per country and announce changes in advance. Treat any rate quoted by a blog, including vendor blogs, as a snapshot that may already be stale.
How do I compare BSP quotes fairly?
Force every quote into the same three lines: Meta's pass-through fees (at cost or marked up?), the provider's platform fee, and per-seat or per-volume software charges. Then model your real monthly mix, template categories included, at both current and 2x volume. The cheapest-looking quote is often the one with the widest per-message markup.