induwara.lk
induwara.lkSri Lanka · AI & Finance

WhatsApp Business + AI Chatbot Cost Calculator

Work out the real monthly cost of a WhatsApp AI chatbot: Meta's per-message fee (Sri Lanka rate card) plus the LLM token cost of the replies, in rupees and dollars. Free, no signup, every rate source cited below.

By Induwara AshinsanaUpdated Jul 4, 2026
WhatsApp bot monthly cost

Sets the seed per-message rates below. Sri Lanka is the default.

Currency
CBSL

Headline totals show in rupees.

WhatsApp messages / month

Template messages by category, plus the per-message rate applied to each.

Free — window open

AI reply layer

Free-form replies inside the service window are free on WhatsApp, but the LLM still charges for tokens.

Input / output price per 1M tokens.

Free-form conversations the bot answers

≈ 798 tokens for a 600-word context

≈ 146 tokens for a 110-word answer

BSP / platform fee (optional)

Business Solution Providers (360dialog, Wati, local aggregators) add a per-message markup and/or a fixed monthly fee. Leave at 0 for Meta-direct Cloud API.

Added on top of every template message

Flat platform subscription, if any

Quick presets
Total monthly cost
Rs 3,325
$10.90
WhatsApp messaging
Rs 1,906
Template + auth fees
AI (LLM) cost
Rs 1,418
Claude Haiku 4.5
Cost per conversation
Rs 0.66
5,000 messages + chats

Cost by layer

WhatsApp messagingRs 1,90657.34%
AI (LLM)Rs 1,41842.66%
BSP / platformRs 0.000%

LLM split — input tokens Rs 732 (2,400,000 tok) vs output tokens Rs 686 (450,000 tok).

WhatsApp cost by category

CategoryMessagesUnit rateLine cost
Marketing
500$0.0125Rs 1,906
Utility
Free — sent inside the open 24-hour service window
1,500FreeRs 0.00
Authentication
0$0.0058Rs 0.00
Service (free-form AI replies)
Free — replies inside the 24-hour customer-service window
3,000FreeRs 0.00
WhatsApp messaging totalRs 1,906

Line-item total verified against the closed-form WhatsApp pricing formula.

Sources cited: Meta WhatsApp Business Platform pricing (per-message model, free service/utility-in-window rules), Anthropic pricing (authoritative for Claude), OpenAI and Google Gemini pricing pages (transcribed, dated 2026-07-04), and the CBSL indicative USD→LKR rate. Full links are in the Sources & references section below. The per-country WhatsApp rates are editable seed estimates pending a direct read of Meta's rate card — paste your exact rate into each $/msg field for a precise bill.

How it works

A WhatsApp AI chatbot bill has two layers that most calculators price only one of. This tool adds them together and shows each layer separately. It models the per-message system Meta moved to on 2025-07-01 — the older conversation-based pricing is gone.

Layer 1 — WhatsApp messaging. Meta bills per message, by category, with country-specific rates. Using M, U, A for the monthly count of marketing, utility and authentication templates and rfor each category's per-message rate:

WhatsApp = rM·M + rA·A + (window open ? 0 : rU·U)

Service messages — your bot's free-form replies inside the 24-hour customer-service window — are always free. Utility templates are free too when sent inside that open window, so they only cost money if the toggle is off. Marketing and authentication templates are always billed. That single free-window rule is the biggest swing in a support-heavy workload.

Layer 2 — the LLM. Only the free-form AI replies call a language model; templates are pre-written and cost no tokens. With R replies, tin and tout average input/output tokens, and pthe model's price per million tokens:

LLM = R · ( (tin / 1,000,000)·pin + (tout / 1,000,000)·pout )

An optional third layer covers Business Solution Providers (360dialog, Wati, local aggregators) that resell the API: a per-message markup b on every template plus a fixed monthly fee F, so BSP = b·(M + U + A) + F. Leave both at zero for the Meta-direct Cloud API. The grand total is WhatsApp + LLM + BSP in USD, converted to rupees at your editable CBSL rate, and cost per conversation is the total divided by all messages and chats (guarded against divide-by-zero).

The WhatsApp line is computed twice — as a sum of per-category line items and again from the closed-form formula above — and the calculator confirms the two agree, the same cross-check discipline used on the income-tax tool.

Worked examples

Nugegoda clothing store — window open

500 marketing · 1,500 utility (free) · 3,000 Claude Haiku replies

  1. WhatsApp: 0.0125×500 + 0.0058×0 + (window open → 0) = $6.25
  2. LLM input: 3,000 × 800 / 1e6 = 2.4M tok × $1 = $2.40
  3. LLM output: 3,000 × 150 / 1e6 = 0.45M tok × $5 = $2.25 → LLM = $4.65
  4. BSP: $0 (Meta-direct)
  5. Total: 6.25 + 4.65 = $10.90 → Rs 3,324.50 at Rs 305
  6. Cost / conversation: 10.90 / 5,000 = $0.00218

Freelancer OTP bot — no AI

10,000 authentication codes · no marketing, no LLM

  1. WhatsApp: 0.0058 × 10,000 = $58.00
  2. LLM: $0 (templates only, no model call)
  3. BSP: $0
  4. Total: $58.00 → Rs 17,690 at Rs 305
  5. Cost / conversation: 58 / 10,000 = $0.0058

Same store — window closed

Utility now billed — shows the free-window swing

  1. WhatsApp: 0.0125×500 + 0.0030×1,500 = 6.25 + 4.50 = $10.75
  2. LLM: $4.65 (unchanged)
  3. Total: 10.75 + 4.65 = $15.40 → Rs 4,697 at Rs 305
  4. The free-window rule alone swings WhatsApp by 0.0030 × 1,500 = $4.50

All figures use Sri Lanka's illustrative rates (rM = $0.0125, rU = $0.0030, rA = $0.0058) and Claude Haiku 4.5 ($1 in / $5 out per 1M). Enter your own rate-card numbers for an exact bill.

Frequently asked questions

Sources & references

LLM prices were last cross-checked on 2026-07-04. The WhatsApp per-country rates are editable seed estimates pending a direct read of Meta's rate card; the arithmetic is exact for whatever rate you enter. Vendor prices move — re-check the linked pages before you commit a budget.

Related tools

Rate this tool
Be the first to rate

Comments & feedback

Spotted a bug or want an improvement? Tell us — our team reviews every comment, and good ideas get built. Comments are public and anonymous.

Found a bug, edge case, or want to suggest an improvement?

Email me at [email protected] — most fixes ship within 24 hours.