WhatsApp Business API Pricing

Update, September 2026: from 1 October 2026 Meta bills WhatsApp per message instead of per 24-hour conversation, and service replies (your answers inside the customer's 24-hour window) stop being free: Meta charges them at the utility rate for the customer's country. The method below still holds. Swap "conversation" for "message", use Meta's current per-message rate, and make sure a payment method is attached to your WhatsApp Business account before 1 October, or replies stop sending.

Two separate companies bill you for one WhatsApp conversation. Almost nobody explains this clearly, and it's the single most common reason a support lead gets surprised by a bill three months into using WhatsApp for customer service. If you've searched "whatsapp business api pricing" and landed on a vague overview, this isn't that.

Meta charges per conversation, based on category and country. Your software vendor, whoever built the chat widget, inbox, and AI replies on top of that API, charges a separate subscription for the platform itself.

Add a "per AI reply" cost on some plans and you have three numbers to track, not one.

Most vendor pricing pages either hide this split entirely or bury it in a footnote, which is exactly how a finance team ends up asking "wait, why are we paying two different companies for WhatsApp" in month two.

This post does the math nobody else bothers to show. We'll build a real, worked example for a 500-conversation month, keep Meta's charge and the subscription charge visible as two separate lines the whole way through, and give you a method to redo the calculation yourself when Meta's rates change. Because they will.

Whatsapp Business Api Pricing Key Takeaways

• Meta bills the conversation. Your vendor bills the software. These are two separate invoices from two separate companies, and any pricing page that blends them into one number is hiding something.
• A realistic 500-conversation month has a Meta cost and a platform cost that need to be added, not confused. We show both lines explicitly using our published Starter ($29), Pro ($99), and Scale ($299) plans.
• Meta's per-conversation rates change over time and vary by country and category. Treat any rate you read (including ours) as a snapshot, and check Meta's own pricing page before you budget.
• Replies your own team types by hand are free on Tendabot and never count against your plan's AI reply limit. Only the AI-generated replies do (500 / 3,000 / 15,000 depending on plan).
• Once you know the method, you can rerun this math for your own conversation volume, country mix, and category split in about five minutes.

Table of Contents

• Why Your WhatsApp Bill Never Matches What You Expected
• Meta Bills the Conversation, Your Software Vendor Bills the Subscription
• What a Real 500-Conversation Month Actually Costs
• What Happens When You Outgrow the Starter Numbers
• Where This Pricing Model Isn't the Right Fit
• How to Run This Math With Your Own Numbers

Why Your WhatsApp Bill Never Matches What You Expected

You budgeted $99 a month for a support tool. Then a second invoice shows up from Meta, and nobody on your team can explain what it's for or why the number changes every month. That's because Meta bills separately from the software, and most onboarding flows never walk you through that split before you connect the Cloud API.

Here's the plain truth nobody puts on their pricing page: WhatsApp Business API pricing is not one number. It's two separate bills, from two separate companies, charging for two separate things.

• Meta bills you per conversation, based on rates set in its own pricing system.
• Your software vendor (the company that built the chat interface, inbox, and AI layer) bills you a flat subscription for the tool itself.

These two charges have nothing to do with each other. Meta doesn't know or care which vendor you use. Your vendor doesn't set or collect Meta's conversation fees. They just show up on your books in the same month, and most guides blur them into a vague "WhatsApp business cost is around $X" statement that helps nobody.

That vagueness is the actual problem. A support lead trying to budget for an AI agent needs to know: what does Meta charge per conversation right now, what does the software cost, and what do both add up to for a real month of volume. Not a hand-wave, not an average. An actual worked number.

That's what the rest of this article does. We'll walk through a realistic month, say 500 conversations, and add up both bills separately, Meta's conversation charge on one line, the software subscription on another. No blending them into a single misleading figure.

** Meta only charges when a conversation happens on WhatsApp, and even then, only under specific conditions. Your subscription, on the other hand, charges based on how many AI-generated replies you use. Those are different meters, tracking different things.

Once you see them side by side, the invoice stops being a mystery.

Meta Bills the Conversation, Your Software Vendor Bills the Subscription

Every WhatsApp Business API setup involves two separate bills, and mixing them up is how buyers get blindsided later. Meta owns the WhatsApp Business Platform itself. It charges per conversation, sorted into categories: marketing, utility, authentication, and service.

Rates differ by country and by category, and Meta adjusts them from time to time, which is exactly why whatsapp business api price lists you find on random comparison sites go stale fast.

That means the only trustworthy number for what Meta charges is Meta's own pricing page. Not this article, not a sales deck, not a chart someone published eight months ago. Rates move. Treat any hard-coded figure you read online as a snapshot, not a promise.

Separate from that, your support platform charges its own subscription. This covers the AI agent, the shared inbox, and the knowledge base that trains it. That fee has nothing to do with what Meta bills per conversation. It's a different vendor, a different invoice, a different line on your budget spreadsheet.

Two vendors, two bills, no overlap. Here's the split in plain terms:

• Meta charges per conversation category, at rates set and changed by Meta, varying by country.
• Your support platform charges a flat monthly subscription for the software that builds and runs the agent, the inbox, and the knowledge base.

Some buyers get frustrated mid-contract because a vendor quoted "$X per conversation" and let them assume that number covered everything, including Meta's cut. It didn't. When a vendor blurs these into one number, you can't tell if you're overpaying the software fee or underestimating the Meta fee. You find out the hard way on your first real invoice.

Showing both separately isn't a technicality, it's how you budget correctly. You need to know your subscription tier, your expected conversation volume, and Meta's current rate for your country and category. Add them and you get your real monthly cost. Skip one and your forecast is wrong before the month even starts.

What a Real 500-Conversation Month Actually Costs

Let's run the actual numbers instead of talking around them. A 500-conversation month is a realistic size for a small ecommerce or B2B support team, so it's worth doing the math properly rather than waving at a "low monthly cost."

Here's the method, and it stays correct even after Meta updates its rates:

• Take your conversation count for the month (500, in this example)
• Multiply by Meta's current per-conversation rate for that category, taken from Meta's official pricing page
• That total is Meta's bill, paid to Meta directly

Here's what that looks like with a real number. [^1] Multiply that by 500 conversations and Meta's bill for the month comes to $18.90, paid directly to Meta, nothing to do with your software subscription.

Swap in whatever your category and country show today, since this exact figure won't hold forever, but the multiplication stays the same.

[^1]: Verified directly against Meta's official WhatsApp Business Platform pricing page on August 5, 2026. Check that page yourself before you finalize a budget, since Meta updates rates without notice and this number is a snapshot, not a permanent fact.

The Software Bill Is Separate

The software bill is a different line item, paid to whoever runs your AI agent. On our pricing page, 500 conversations lines up closely with the Starter plan at $29/mo, which includes 500 AI replies. That $29 covers the platform, the shared inbox, the WhatsApp connection, and the AI itself. It has nothing to do with what Meta charges per conversation.

So your real monthly cost has two separate totals, not one blended number:

• Meta's bill: 500 x $0.0378 (US utility rate, verified against Meta's pricing page on August 5, 2026) = $18.90
• Tendabot's bill: $29/mo flat, for the Starter plan's 500 AI replies

Add those two lines and this example month runs $47.90 total, Meta's $18.90 plus Tendabot's $29, kept apart the whole way through. Blending that into a single "cost per conversation" figure is where most whatsapp api pricing guides go wrong. Keep them apart, because they're billed by different companies on different schedules.

One more detail worth knowing: if your team jumps in and types a reply by hand instead of letting the AI answer, that reply is free and doesn't count against your 500-reply limit. Only AI-generated replies get metered. Run out of AI replies and the bot pauses politely rather than quietly racking up an overage bill.

What Happens When You Outgrow the Starter Numbers

Five hundred conversations a month is exactly what the Starter plan is built for, at least at first. As WhatsApp becomes the channel customers actually prefer over email or the website widget, conversation volume climbs, and 500 AI replies stops covering a full month. That's not a failure of planning. It's just what happens when a side channel turns into the main one.

Here's how the plans map to volume:

• Starter ($29/mo): 500 AI replies, fine for early testing or low-volume support
• Pro ($99/mo): 3,000 AI replies, the range where most growing teams land
• Scale ($299/mo): 15,000 AI replies, built for teams running WhatsApp as a primary channel

Meta's per-conversation fee does not care which of these plans you're on. It scales the same way whether you're on Starter or Scale, because it's billed by Meta directly, separate from your subscription. Move from 500 conversations to 3,000, and your Meta bill roughly grows with it.

Your subscription only changes if you upgrade plans. These are two independent lines, not one blended number, and treating them that way is the only way the math stays honest.

There's a safety net worth knowing about. If your AI reply allowance runs out mid-cycle, we pause the AI instead of quietly billing you for overage. Say your reply volume hits 3,200 in a 3,000-limit Pro month. The AI pauses at 3,000, your team keeps answering by hand at no charge, and your subscription bill stays flat at $99, no overage line, no surprise charge.

Scale adds one more lever. It unlocks BYOK across all 41 models from 11 providers, so teams with real volume can plug in their own API keys and pay each provider's actual token cost with no markup added on top. For a team running 10,000+ AI replies a month, that control over model costs often matters more than the flat subscription fee itself.

Where This Pricing Model Isn’t the Right Fit

Two separate bills are not for everyone, and it's better to say that plainly than watch someone sign up and feel misled later. Our live channels today are website chat, WhatsApp, and email. If a business needs a channel we don't run yet, that's a real gap, and no amount of pricing math fixes it. Check every feature before you commit, so there's no surprise later.

The math in this post also assumes inbound support conversations, not outbound blasts.

If a team runs high-volume marketing campaigns on WhatsApp (order confirmations, promos, cart reminders pushed to thousands of numbers), Meta's marketing-category rate applies, and it runs higher than the utility or service rates used in our worked example above. No software subscription changes that.

The Meta invoice will dominate the total bill regardless of which platform sits on top of it.

There's also a simpler reason this setup won't fit some teams: they just don't want two invoices.

** They act as the Business Solution Provider, mark up the per-conversation rate, and hand you a single monthly charge with no separate Meta line item to track. It's simpler to read on a bank statement. Markups vary a lot by provider.

Some Business Solution Providers publish their per-conversation multiplier directly on their pricing page, others fold it into a flat number and never show the underlying Meta rate at all.

Check the provider's own pricing page for that multiplier before you assume a bundled rate saves you money, because whatever it is, it applies to every conversation you send, all year, not just the first one.

Our no-markup approach rewards a different kind of buyer:

• Ops leads who don't mind checking two numbers once a month
• Founders who want to see exactly what Meta charges versus what the software costs
• Teams who'd rather redo the arithmetic occasionally than pay a permanent markup to avoid it

If tracking a second invoice sounds like a hassle rather than a five-minute check, a bundled provider will probably feel more comfortable, even at a higher long-run cost. If you want to see how the split actually looks inside a real inbox, book a live demo instead of taking our word for it.

How to Run This Math With Your Own Numbers

Every whatsapp business pricing estimate follows the same four steps, no matter what Meta charges this month or next. Run this once, and you can rerun it every time your volume changes or Meta updates its rate card.

Step 1: Pull your real conversation count. Look at last month's WhatsApp chats, not a guess. If you don't have this yet, count the number of unique customers who messaged you in a 24-hour window.

Step 2: Check Meta's current rate. Rates vary by conversation category (service, marketing, utility) and by the customer's country. Meta publishes the live numbers at business.. Verified against that page on August 5, 2026, a US utility conversation runs $0.0378, so 500 conversations at that rate comes to $18.90, but pull the live figure for your category and country before you finalize a budget.

Step 3: Pick your plan separately. This is a different number, based on AI reply volume, not conversation count. Check our current pricing breakdown:

• Starter, $29/month, 500 AI replies
• Pro, $99/month, 3,000 AI replies
• Scale, $299/month, 15,000 AI replies

Replies your own team types by hand never count against these limits. They are free, always.

Step 4: Add the two totals, but keep them as two lines. Meta's conversation fee plus your subscription equals your real monthly cost. Do not blend them into one number in your spreadsheet. You'll need to adjust each side independently when Meta changes rates or when your reply volume grows.

This approach holds up regardless of what Meta charges next quarter, because you're pulling the rate fresh each time rather than trusting a number that gets stale. Bundling a variable, third-party cost into one flat number makes budgeting harder, not easier, because you can't tell which side moved when the total changes. Splitting the two lines, as tedious as it sounds, is the more honest habit.

One AI agent handles your website chat, WhatsApp and email from a shared inbox, running on any of 41 models across 11 providers, billed at the real token cost with no markup. When your plan's replies run out, the bot pauses politely instead of quietly generating an overage bill.

See the exact plan breakdowns and current AI reply limits before you estimate your own 500-conversation month with whatsapp business api pricing in mind.

Start free — 300 credits, no card

Frequently Asked Questions

Does Meta charge per message or per conversation?

Until 30 September 2026 Meta billed per 24-hour conversation. From 1 October 2026 it bills per message delivered, by category (marketing, utility, authentication) and by the customer's country, and service replies inside the customer's 24-hour window are charged at the utility rate instead of being free. Check Meta's pricing page for the live per-message rate before you budget.

Who do I actually pay, Meta or the platform?

Both, separately. Meta invoices you directly for conversation charges through your WhatsApp Business Platform billing setup, and we invoice you separately for the software subscription (Starter, Pro, or Scale). There is no single combined invoice, which is exactly why this article keeps the two numbers apart instead of blending them.

What happens if I go over my AI reply limit mid-month?

The bot pauses AI-generated replies politely instead of billing you for overage. Your team can keep answering customers by hand in the shared inbox at no extra cost, since human replies are always free and never metered, while you decide whether to upgrade to the next plan.

Does Meta’s rate depend on which plan I choose?

No. Meta's per-conversation rate is set by Meta based on country and category, completely independent of whether you're on Starter, Pro, or Scale. Upgrading your plan changes your AI reply allowance and monthly subscription fee, but it has no effect on what Meta bills you for conversations.

Can I estimate my WhatsApp costs before switching from another provider?

Yes, using the four-step method in this article: pull your real conversation count from last month, check Meta's current rate for your category and country, pick the plan matching your AI reply volume, then add the two totals as separate lines. This works whether you're moving off a bundled provider or setting up WhatsApp Business API for the first time.