Guides
The cheapest AI chatbot, measured per message
Monthly prices hide the real cost. We divided each plan by the replies it actually buys, using each vendor's own pricing page, checked 18 October 2026.
₹2,999/month · 5,000 messages · no per-seat or per-resolution fee
A chatbot that costs $40 a month is not cheap if it buys 700 replies. The number that decides your bill is the price of one answered message, and almost no pricing page prints it. So we worked it out for eight tools.
Where a vendor sells credits, we used its cheapest model, which flatters them: a premium model multiplies their figure several times over. Eteon has one price per message, whatever answers.
Cost per reply, side by side
Plan price divided by the replies it includes, cheapest first. The last two rows bill per conversation and per resolution, so they cannot be divided the same way and are shown in their own unit.
| Tool | Plan | Price a month | Replies included | Cost per reply |
|---|---|---|---|---|
| Eteon | Enterprise | ₹4,599 (about $53) | 20,000 | $0.0026 |
| Chatling | Scale | $295 | 70,000 | $0.0042 |
| Eteon | Pro | ₹3,999 (about $46) | 10,000 | $0.0046 |
| Eteon | Starter | ₹2,999 (about $34) | 5,000 | $0.0069 |
| Chatling | Starter | $25 | 3,500 | $0.0071 |
| Boei | Starter | $25 | 2,000 | $0.0125 |
| SiteGPT | Starter | $59 | 4,000 | $0.0147 |
| FastBots | Essential | $39 | 2,000 | $0.0195 |
| Chatbase | Hobby | $40 | 700 | $0.0571 |
| Tidio | Lyro AI add-on | from $32.50 | 50 conversations | $0.65 per conversation |
| Intercom | Fin AI agent | from $29 per seat | billed per outcome | $0.99 per resolution |
Monthly billing, from each vendor's pricing page on 18 October 2026. Eteon is priced in rupees and converted at ₹87 to the dollar. Chatling sells credits and its cheapest models use half a credit per reply, so 35,000 credits buy 70,000 replies and 1,750 buy 3,500. Boei's 2,000 credits are counted as one reply each; premium models use more.
The winner on price: Eteon
Eteon Enterprise is ₹4,599 a month for 20,000 messages: ₹0.23 a message, about a quarter of a US cent. Pro is ₹0.40, and Starter, the entry plan at ₹2,999 for 5,000 messages, is ₹0.60, still under a cent.
Measured as value for money, replies bought per dollar, nothing else in the table comes close at the top plan: the next cheapest, Chatling Scale, costs about 60% more per reply.
There is no per-seat fee, no per-resolution fee and no model multiplier, so the figure in the table is the figure on the invoice.
Cheap only matters if it sells. Eteon was built as a salesperson first: it answers from your own pages and catalogue, handles the objection that stops a purchase, confirms cash-on-delivery orders on WhatsApp before they ship, and answers "where is my order?" at any hour.
Where the cheap plans hide the cost
Most low-priced plans sell credits, not messages, and a credit is only a message on the vendor's cheapest model. Choose a stronger model and the same plan buys a fraction of the replies: FastBots charges 5 or 10 credits a reply on its larger models, Chatling up to 25.
The headline allowance can also be small. Chatbase Hobby is $40 a month for 700 credits. Top-ups are $40 per 1,000 credits, and removing the "Powered by Chatbase" line costs $1,188 a year on top.
Eteon counts one visitor message and the reply to it as one message, on every plan. The small Eteon mark on Starter comes off from Pro upwards, at ₹3,999 a month, with nothing extra to buy.
Per seat and per resolution: the expensive model
Support suites price AI the way they price people. Intercom charges from $29 a seat a month, then $0.99 for every conversation its Fin agent resolves, so a thousand resolved conversations is about $990 before the seats.
Tidio's Lyro AI starts at $32.50 a month for 50 conversations, which is 65 cents a conversation. Both models were designed for human teams, and both make a good month more expensive.
Yearly billing changes one row
Eteon bills monthly only, so the table compares monthly prices. Several vendors discount yearly plans by 15 to 35 percent, and that moves one comparison: Chatling Starter billed yearly is $21 a month, or $0.0060 a reply, a little below Eteon Starter.
It does not move the top of the table. Eteon Enterprise at $0.0026 a reply on monthly billing is still below every yearly price we found, the closest being Chatling Scale at $0.0035, and nothing is locked in for twelve months.
When the cheapest is not the right pick
If you will send fewer than a couple of hundred messages a month, a free plan costs nothing: Chatling's gives 200 credits a month, Chatbase's 50. That is enough to test a bot, not to run a shop on.
If you need one inbox across fifty-odd channels, Boei is built for that. If you run a support team that needs ticketing, SLAs and seats, a helpdesk like Intercom or Zendesk is the right tool, and the higher price buys it.
If you want a bot on your website that sells, answers from your own pages and costs less per message than anything else here, that is what Eteon is for.
How we measured
Prices come from each vendor's own pricing page, checked on 18 October 2026, on monthly billing. We divided the plan price by the replies it includes. Where a vendor sells credits, we used the credits one reply costs on its cheapest model, which is the most favourable reading for them.
Eteon's prices are in rupees and converted at ₹87 to the dollar. A message is one visitor message and the reply to it; your own testing in the dashboard does not count. Prices change, so the date above is the date these were true.
Questions
Frequently asked questions
What is the cheapest AI chatbot?
Of the eight tools we priced on 18 October 2026, Eteon has the lowest cost per message: ₹0.23 (about $0.0026) on Enterprise and ₹0.60 (about $0.0069) on Starter, with no per-seat or per-resolution fee.
Is there a free AI chatbot?
Several tools have free plans, but they are small: Chatling gives 200 credits a month, Chatbase 50, and Tidio 50 AI conversations in total. They are fine for testing. Eteon gives 30 days free with 500 messages instead.
Why is Eteon cheaper per message?
Two reasons. Per-seat and per-resolution meters were designed for human support teams and price AI as if it were a salary. And we rebuilt the prompt around what actually gets billed per turn, which cut the cost of a conversation by roughly six times; that saving is in the price.
Does cheaper mean worse answers?
Not by itself. Answer quality comes from what the bot is allowed to say. Eteon answers only from your own site and catalogue, and any price that does not appear there is removed from the reply before the shopper sees it.
What counts as a message?
One message from a visitor and the reply to it. Your own testing in the dashboard does not count against the allowance.
Build one on your own site.
Enter a URL, talk to the bot it produces, and decide from that rather than from this page. Nothing is charged for 30 days.