The Live Chat Software Market, Grouped by Meter

Live chat is the small, fast channel on a website. A visitor types a question, an answer appears in the same window, the visitor leaves. Underneath that simple flow sit pricing decisions that look technical and are actually commercial: what unit is the vendor going to count, and what happens when that unit doubles.

This page is the side-by-side of the five meters in common use, with each vendor's published prices written exactly as they appear on its pricing page on 4 September 2026. The aim is to show the shape of each ladder, what it costs at three realistic sizes, and which kind of business each meter is really built for.

Five products, five meters

Each of the products below is sold on a different unit. That single fact does more to predict your annual bill than any feature comparison.

Meters are taken from each vendor's own pricing page, checked 4 September 2026. The "who it fits" column is the typical buyer for that meter, not a judgement on the product. We make Asyntai, so read the last row as a disclosure.

LiveChat: priced per agent

LiveChat sells three plans and the unit is the named human who logs in. Annual billing is the cheaper column; paying monthly adds about $10 per person to each tier.

We make Asyntai, so read that row as a disclosure. The Asyntai Starter plan is $39 a month for 2,500 messages, regardless of how many named humans look at the inbox.

A team of four on the Team plan paid annually works out to $196 a month, $2,352 a year. The same team paying monthly costs $236 a month, $2,832 a year. The published ladder rewards commitment: annual billing is the cheaper column.

Per-agent pricing is the simplest to budget for, because you already know the headcount. It is also the most punishing when volume spikes. Adding one seasonal helper for a month costs the same $49 as keeping a permanent team member, because there is no short-term rate on the published ladder.

Zendesk: priced per agent, with add-ons stacked on top

Zendesk's live chat is the chat product inside its wider Support Suite. The agent price is the base layer. The AI and routing products are billed on top, per agent.

We make Asyntai, so read that row as a disclosure. The Asyntai Starter plan is $39 a month for 2,500 messages with no per-agent add-ons on top.

Note that Suite Enterprise is sold by quote, not on the self-serve ladder. If your buyer is enterprise, the published price list does not apply to you.

That is the trap with stacked add-ons. The headline plan looks modest. The bill grows by an extra line per agent for each capability, and you find out how much AI actually costs by adding up the column. Per-agent pricing with per-agent add-ons makes the bill scale linearly with the size of your team, which is fine when the team is fixed and awkward when it changes.

Intercom: priced per seat, with AI priced per resolved conversation

Intercom splits the bill into two different meters, and that split is the interesting design decision. The human product is sold by seat. The AI product, Fin, is sold by resolved conversation.

We make Asyntai, so read that row as a disclosure. Asyntai does not charge per seat and does not charge per resolution; the unit is the message.

The AI rate is $0.99 per resolved outcome, and it is charged once per conversation. "Resolved" here means Fin handled the customer's question without handing off to a human. If the chat escalates, the $0.99 is still charged for the work Fin did before the handoff.

Two things follow. First, the AI cost is decoupled from the headcount. Add a hundred visitors and the seat bill does not move, but the Fin bill moves by the number of resolved chats. Second, the AI cost is decoupled from whether your team is good. If Fin resolves 2,000 conversations a month instead of 800, the line item triples. The price is per outcome, so the bill rewards volume.

Crisp: priced per workspace

Crisp's meter is the workspace, not the seat. Each plan has a fixed number of seats bundled in, and extras are billed per agent at $10 each.

We make Asyntai. The $39 Starter plan has no seat cap; the unit is the message, not the human.

Workspace pricing is the right shape for agencies and small groups running several brands on one bill. One Mini workspace covers four people across as many client sites as you want. Add a fifth person for a project and the line is $10 a month, not a plan upgrade.

Notice the jump from Essentials to Plus. Essentials covers 10 seats for $95. Plus covers 20-plus seats for $295. If you cross that threshold and do not need the Plus feature set, the per-seat cost in Plus is higher than the per-seat cost in Essentials, even though you are paying three times the monthly fee. Workspace pricing has cliffs, and the cliff between Essentials and Plus is a sharp one.

Tidio: priced per conversation

Tidio is the only product in this group that does not charge for the human at all. The unit is the conversation, defined by the visitor side: each new chat from a unique visitor is one conversation.

We make Asyntai. The Starter plan is $39 for 2,500 messages, which is roughly the monthly volume Tidio charges about ten times as much to reach.

Per-conversation pricing is the right shape for a single store or a small site where nobody is full-time on chat. The free 50 covers most early sites. Starter at $24.17 covers 100 conversations, which is enough for many small businesses. Growth starts at $49.17 for 250 conversations and rises with volume.

The thing to know about per-conversation pricing is that the bill moves with the visitor, not with your team. A marketing campaign that triples chat volume triples the bill. That is good when your team is small and you only want to pay for what you use, and bad when your site gets an unexpected spike.

Reading the ladders together

Group the five meters and the pattern is clear. The vendors split into three camps on what they are selling.

The last row is Asyntai. We make it, so read it as a disclosure. Per-message is the fifth meter in this comparison and the only one where neither the people nor the visitor count appear on the invoice.

Per-agent and per-seat pricing assumes the team is the fixed thing and the visitor traffic flexes around them. Per-workspace pricing assumes the brands are the fixed thing and the people flex around them. Per-conversation pricing assumes the visitor is the fixed thing and everything else flexes around them. Each one is right for a different business shape, and wrong for the others.

Which meter fits which business

This is the practical question, and the answer is sharper than the feature comparisons usually admit.

The last row is Asyntai. We make it. Read it as a disclosure. The shape that suits a small team running chat on a single site is the shape the per-message meter was built for.

Three worked examples at realistic sizes

Pick three business shapes and run each through the published ladders. Use only the numbers in this article.

A four-person support team on one site

Tidio is omitted here because its meter is the conversation, not the team. A four-person team answering thousands of chats a month would pay Tidio's Growth band, which sits above the four-agent per-seat ladders; the per-conversation shape is wrong for that workload.

An agency with eight people across twelve client sites

This is where per-workspace pricing pulls ahead. Crisp charges for the seats and the workspaces, not for the sites. The agency runs twelve client sites on one Mini workspace with eight seats for $85 a month. LiveChat bills per agent regardless of how many sites that agent covers, so the agency pays for the people twice over once you account for the brands.

A SaaS that wants AI to handle most chats and humans to handle the rest

Per-resolution AI pricing rewards the vendor when the AI works, because each resolved conversation is a metered event. If Fin resolves half your chats, you pay $0.99 for each one. If Fin resolves none of them, you pay the seat rate only. That is the right meter if you believe AI is doing most of the work, and the wrong meter if you expect humans to handle most conversations.

Where Asyntai sits in this market

Asyntai sits at the bottom of the meter list above, on a unit of its own. The unit is the message, and the named-human count is not on the invoice.

The four self-serve tiers are Free, Starter, Standard and Pro. Enterprise is sold on a signed order form with a negotiated message allowance and is not on the public pricing table.

The shape that matters is the column, not the row. A small business with one person checking the inbox once a day pays $39 a month for 2,500 messages, the same as a five-person team in the same situation, because the named-human count is not part of the price. The bill moves with the messages, the same unit Tidio uses, but the published ladder is flatter: 2,500 messages for $39, 15,000 for $139, 50,000 for $449.

Per-message pricing is closest to per-conversation pricing, with one difference worth knowing. Conversations are a vendor-defined unit, and the definition changes between products. Messages are easier to count on your own site before you sign anything, which is what makes the per-message meter the easiest to budget against.

The mistake most buyers make

The common mistake is to compare the headline price of each plan rather than the meter. LiveChat's $19 Starter and Zendesk's $19 Support Team look like the same offer, and they are not. LiveChat's $19 is for the chat product on its own. Zendesk's $19 is for the Support Team plan, which is a thinner bundle inside a wider suite, and the moment you add Copilot or Workforce Management the per-agent add-on column kicks in.

The other mistake is to fixate on the per-unit rate without modelling the units. A $0.99 per resolved conversation looks cheap next to a $19 per agent per month. Multiply it by 800 conversations and the AI line is $792, which is several agent-months of base fee. The headline rate is not the bill; the bill is the rate times the units you actually consume.

Third mistake: assuming your volume will stay flat. Per-conversation pricing punishes growth. Per-agent pricing punishes adding seats for a project. Per-workspace pricing has cliffs at the bundle boundaries. Per-message pricing is the flattest of the four, and the one where a doubling of volume costs the next published bundle rather than a custom quote.

What to measure on your own site before you sign

Three numbers decide which meter is right. Pull them from your analytics before you talk to any vendor.

1. Monthly chat volume. Open your current chat tool or your email inbox for support and count the live chats. If the number is under 500 a month, per-conversation pricing is competitive. If it is over 2,000 a month, per-conversation pricing has lost the comparison before it starts.
2. Named humans who will log in. Count the team members who need to see the inbox, not the company headcount. A per-agent or per-seat meter is cheaper when this number is small and stable, and more expensive when it is large or seasonal.
3. Number of sites or brands on one bill. If you run several sites or client brands, per-workspace pricing saves the most. If you run one site, per-workspace pricing is a more expensive way to do per-agent pricing.

Match those three numbers to the meter and the comparison becomes shorter. A team of four with 1,500 chats a month on one site is in per-agent or per-message territory. A team of four with 1,500 chats a month across twelve sites is in per-workspace territory. A team of two with 200 chats a month on one site is in per-conversation territory, and probably should be on the free tiers of several products to test which UI the team can stand to use — the kind of shortlist a best live chat for small business roundup works through in detail.

How to read this page a month from now

Pricing pages move. The numbers above are taken from each vendor's pricing page on 4 September 2026. LiveChat's annual versus monthly split, Zendesk's per-agent add-ons, Intercom's per-resolved Fin rate, Crisp's workspace bundles and Tidio's conversation bands are all published prices at that date. Re-check them before you sign, because the published ladder is the only thing that matters for budgeting.

The shape of the market, though, is slower to change. Five products and five meters is the published reality today, and the choice between them is a choice between who the product is priced for. Pick the meter first, then the vendor. The other way around is how most people end up paying for capacity they did not need.