Usage Based Chatbot Pricing: Find the One Number Being Metered, Then Forecast Your Bill in Three Busy Months (Free Meter Chooser)
Usage based chatbot pricing sounds fair, because you pay for what you use. The catch is the word “use”. It can mean visitors, conversations, messages, AI replies or resolved tickets, and each one grows differently. This guide shows how to find the single metered number, forecast it for your busiest months and pick between metered and flat plans.
⚡ In short
- Every metered plan counts one thing. Find out which, because it decides how your bill grows.
- Forecast with your busiest three months, not your average month.
- Prefer a meter you can predict from analytics, such as visitors, over one you can’t, such as AI tokens.
- Use the decision tree below to see whether metered or flat pricing suits you.

The basics
What usage based chatbot pricing really measures
With a flat plan you pay the same each month. With metered pricing the bill follows a counter. Vendors choose very different counters, so two plans that both say “pay as you use” can behave nothing alike.
Common meters
- Visitors. Counts people who open your site or trigger the chat. Easy to predict from your analytics.
- Conversations. Counts chats started. Closer to real work, but a busy day can surprise you.
- Messages. Counts every line sent, so long chats cost more.
- AI replies. Counts answers the bot writes. Grows with how chatty the bot is.
- Resolved tickets. Counts problems closed. Fair, but hard to forecast.
The honest question for any plan is: what exactly increments the counter, and can I see it before the invoice?
Good and bad
When metering works in your favour
Metered billing is a good deal when your traffic is small or lumpy. A seasonal shop that is quiet for eight months and busy for four doesn’t want to pay the same every month.
It is a bad deal when the counter is hard to see or hard to control. If you can’t tell which pages push the number up, or the plan charges automatically when you pass a limit, you are buying uncertainty.
Three signs of a friendly meter:
- The count is visible in your dashboard, in real time.
- You can set a ceiling, so you cannot overspend by accident.
- Going over the limit slows the service rather than charging your card.
If a vendor can’t answer those three points in writing, treat the pricing as unpredictable, however cheap the unit price looks.
Try it
A meter chooser: metered or flat?
Follow the path that matches your traffic and your tolerance for surprises.
Forecasting
Forecast usage based chatbot pricing with three busy months
The unit price tells you almost nothing until you multiply it by a realistic count. Use this method.
Step 1: find your busiest three months
Open your website analytics and write down monthly visitors for the last twelve months. Pick the top three. A sale, an exam result or a festival often creates the peak.
Step 2: turn visitors into the vendor’s meter
Not every visitor chats. If the plan meters visitors, use the number directly. If it meters conversations or AI replies, estimate how many visitors start a chat (try a pilot first, then use your real rate) and how many messages a chat contains.
Step 3: price each month, then add a buffer
Multiply, then add 30% for growth. The result is the number you should compare between vendors, not the headline rate.
Traps
Four traps hiding in usage based chatbot pricing pages
Read the small print for these.
- Overage rates higher than the base rate. The first 1,000 units are cheap and the next are not.
- Automatic top-ups. Your card is charged the moment you pass a threshold.
- Unequal units. One “AI reply” may be counted as several if the answer is long.
- Minimum commitments. A “pay as you go” plan with a monthly floor is a subscription in disguise.
Our guide to website chatbot subscriptions versus pay-as-you-go works through a seasonal year, and the chatbot SaaS pricing article explains how vendors build these plans.
Worked example
A worked forecast with illustrative numbers
These numbers are made up to show the method, not real data.
Imagine a small online shop with 600 visitors in an average month. Its three busiest months, around a festival, reach 1,800, 2,200 and 1,500 visitors. Add 30% growth to the peak and plan for roughly 2,900.
If the plan meters visitors
You can compare 2,900 directly with each plan’s allowance. Count how many packs or tiers you need for that month, then multiply by the number of peak months. The other months cost far less.
If the plan meters AI replies
Suppose one visitor in ten starts a chat and each chat gets six bot replies. At 2,900 visitors that is about 290 chats and about 1,740 replies. A plan that includes 1,000 replies would run out in the peak.
What the example shows
The same traffic gives very different bills depending on the meter. Always do the multiplication before you trust a unit price, and always do it for the peak month.
Our approach
How LiveChatBot meters usage
LiveChatBot uses a visitor meter and keeps it simple.
Each website gets 100 free visitors a month. A visitor is counted once per month, when they trigger a new-visitor alert. Beyond that you can buy one-time visitor packs: 100 for ₹99, 500 for ₹299 and 1,000 for ₹499. Packs never expire, so a quiet month doesn’t waste them. Two websites are free; an extra website is ₹199 for 6 months. There are no per-seat charges.
If you pass the limit, tracking and alerts continue, and chat pauses only for new over-limit visitors until you buy a pack or the month resets. Nothing is charged automatically.
The AI Agent is separate. It has a small free monthly allowance per website, then AI packs, and the free keyword bot answers from saved replies when the allowance is used up. See the pricing page for the current figures.
Honest limits
What this model does not do
A visitor pack is not unlimited. If your site gets tens of thousands of visitors a month, count the packs before you decide. A flat enterprise plan may suit that traffic better.
There is no live “spend cap” slider to set, because nothing is billed automatically. Instead the cap is built in: chat pauses for new over-limit visitors. That protects your wallet but means you must watch your count in busy periods.
Before you commit anywhere, run a 30-day pilot on the free plan and watch the visitor count. It costs nothing and gives you the real number for your forecast. The AI chatbot pricing guide has a worksheet for comparing quotes side by side.
Frequently asked questions
What is usage based chatbot pricing?
A pricing model where your bill follows a counter, such as visitors, conversations, messages or AI replies, instead of a fixed monthly fee.
Is metered pricing cheaper than a flat plan?
Often for small or seasonal sites, rarely for steady high traffic. Forecast your busiest three months before deciding.
What does LiveChatBot count?
Visitors per website. Each visitor is counted once per month, with 100 free per website and one-time packs after that.
Does LiveChatBot charge automatically when I pass the limit?
No. Chat pauses for new over-limit visitors until you buy a pack or the month resets.
Do packs expire?
No. Visitor packs never expire, which suits uneven traffic.
Keep reading
Related guides
Pay as You Go Chatbot: Prepaid Packs or Metered Billing? Follow One Pack From Purchase to Empty (Free Step-by-Step Timeline)
Chatbot Pricing for Small Websites: Why a Few Hundred Visitors a Month Changes Which Price Shape Fits (Free Price-Shape Quiz)
Low Cost Chatbot Hosting: Shared Hosting, a Small VPS, Serverless or a Hosted Widget? Score the Four Routes for Your Own Priorities (Free Route Scorecard)
Try a meter you can see
Two websites free, 100 free visitors per website each month and one-time packs that never expire. Start free on livechatbot.in.