Facebook Chat Plugin Alternative: Where the Conversation Lives

The Facebook customer chat plugin was deprecated in 2024. A lot of small sites still rely on what replaced it, the Messenger experience embedded on a page, or a manual link. Both routes send the visitor off your domain into a Messenger window. That single fact is the reason to read this page rather than keep the plugin.

Most comparison posts describe the chat widget. The interesting question is what happens to the transcript after the visitor closes the tab.

What the plugin actually does

The Facebook chat plugin is a small JavaScript snippet you paste into your site. When a visitor clicks it, Messenger opens. The conversation is on Facebook's domain, against your Facebook Page, and your visitor needs a Messenger account (in practice, a Facebook login) to read it back later. You see it from your Page inbox or, with the right setup, in a third party inbox tool that connects to it.

Three consequences fall out of this:

1. The visitor leaves your site to read the thread. There is no record on the page they came from.
2. If the visitor is not logged in to Facebook or Messenger on the same device, they cannot pick up the thread tomorrow.
3. The transcript is yours to read from a connected inbox, but it is not stored against the page, the campaign, or the visitor's session on your site.

This is the trade-off. You pay nothing for the plugin. You pay for it in ownership of the conversation.

Published price, lowest paid tier

Each bar is the cheapest non-free plan that vendor publishes. Scale is $0 to $300 per month.

Tidio

$24.17

tawk.to

$19 (no branding)

LiveChat

$19 per agent

Asyntai

$39

Crisp

$45

Scale: $0 to $300 / month

Where the transcript lives

Facebook plugin keeps it on Facebook. The visitor must be logged in to read it back. Crisp, Tidio, LiveChat and Asyntai keep the thread on your site.

What the meter charges

LiveChat is per agent. Crisp is per workspace seat. Tidio is per conversation. Asyntai is per message. The same dollar number means different things.

What "an alternative" actually means

Three different things are called Facebook chat plugin alternatives, and they are not the same product.

The first two rows keep the visitor on your site and the transcript readable by you. The third row is the same trade-off as the Facebook plugin, in a different app. If the goal is to move off Facebook, the third row is not the answer.

The published prices, side by side

Every figure below comes from the vendor's own pricing page on the date shown. Where the vendor charges per agent, that is the line printed. Where it charges per workspace or per conversation, that is the line printed. Mixing the two meters in one column hides the real comparison, so the table is in two halves.

Per workspace or per branding tier

We make Asyntai, so read that row as a disclosure.

tawk.to's $19 is not really a paid tier. It is a badge removal fee. The chat, the seats, the history, all of it stays free.

Per conversation or per agent tiers

We make Asyntai, so read that row as a disclosure.

Tidio's conversation meter and Asyntai's message meter look like the same thing and they are not. A "conversation" in Tidio's pricing is a chat opened by a visitor, and it counts even if the bot answers it in two exchanges. Asyntai's meter is on messages, which means a single visitor's thread of a dozen turns uses a dozen messages.

Two things to notice in this table. Tidio and LiveChat both have a clear jump between the entry paid tier and the third. LiveChat's $19 to $79 ladder is a per-agent price, so the bill is seats times dollars. Asyntai and Tidio are volume meters, so the bill is conversations or messages, not people.

What the free tiers actually cap at

Free is the right place to start. Every vendor on this page has one. The cap is what matters, because the cap is the moment the free tool stops being free and you either upgrade or leave.

"Free cap" means the published ceiling in the vendor's own pricing. LiveChat does not publish a free tier, so it does not appear in this table.

tawk.to is the unusual case. The free tier is not a trial, it is the product, and the paid line is a branding fee rather than a feature unlock. If a tawk.to badge on your widget is acceptable, free is genuinely the answer.

What you lose when you leave the Facebook plugin

This is the part most comparison posts skip. The Facebook plugin has a real upside: many visitors already have Messenger open on their phone, and a tap into a known app feels lighter than filling out a form. Three things change when you move off it.

1. No more "continue on phone". A visitor who opens the chat on a laptop cannot carry it into Messenger on their phone unless the tool you pick also has a mobile app. LiveChat, Crisp and Tidio do. Asyntai is a web widget, so the thread is on whatever device opened it.
2. You own the transcript. With a hosted widget, every chat sits on your vendor's servers and is exportable from there. You decide how long you keep it. With the Facebook plugin, the visitor's side of the thread lives on Facebook's servers, and you only see what is mirrored into your Page inbox or a connected inbox.
3. The visitor's identity is not assumed. The Facebook plugin already knows who the visitor is on Facebook, which is convenient when it works and a privacy question when it does not. Hosted widgets ask for an email or nothing at all.

For most small sites the third point is the trade that matters. The Facebook plugin implicitly relies on a Facebook login. A visitor who is not logged in to Facebook on that device cannot pick up the thread tomorrow.

Where the transcript ends up

The Facebook plugin sends the visitor's side of the conversation into Facebook. You read it from your Facebook Page inbox, or from a tool that connects to it. Your own site sees nothing.

A hosted widget stores the chat against your workspace on the vendor's servers. You read it from the vendor's dashboard, and you can usually pipe it to a CRM, email, or webhook. Your own site's database sees nothing unless you write that integration yourself.

Asyntai stores the chat against your account on Asyntai's servers, with per-website scoping and configurable retention with auto-delete. The thread is yours to read from your dashboard. Your site's database sees nothing unless you wire that in.

The interesting question is who can read the thread in twelve months. With the Facebook plugin, the visitor's side of it is in Facebook's logs. With any hosted widget, it is on the vendor's servers under your account. With anything self-hosted, it is on your own database, and the question is whether your CMS, your host and your backup policy will still be there.

The price that changes the answer

Free is genuinely free for the Facebook plugin. There is no bill. The price of a hosted widget is what you pay above free. The price that usually decides the choice is the price at which the free tier stops being enough.

For Tidio that is conversation 51, which is also the moment a paying tier becomes $24.17 a month. For Asyntai it is message 101, which is also the moment a paying tier becomes $39 a month. For Crisp it is contact 101, which is also the moment a paying tier becomes $45 a month. For tawk.to, the cap is "never, unless the badge bothers you", and the badge removal is $19 a month.

For LiveChat, there is no free tier. The cheapest entry is $19 per agent per month on an annual bill, or $25 per agent per month monthly. A team of three on Starter is $57 a month on the annual rate.

Notice what is missing. None of the figures above say what happens at scale on the cheap tier. The published next step is the question.

What the next paid tier costs

Every vendor on this page raises its price sharply at the second paid tier. The size of the jump is the price of outgrowing the entry plan, and it is worth printing rather than guessing.

The Asyntai jump looks large because Standard is a 15,000-message tier, which is six times the size of Starter. The dollar-per-message rate falls across the ladder; the headline number rises. We make Asyntai, so read that row as a disclosure.

Per-agent jumps compound with team size. A team of five on LiveChat pays $245 a month at Team, and $395 a month at Business. The same team on Tidio's Growth is $49.17 a month until they pass 250 conversations a month, at which point the bill climbs again.

What changes when the transcript is yours

This is the part the article has been building toward. The Facebook plugin is free and it costs you ownership of the conversation. A hosted widget is paid and it gives you ownership of the conversation, but it does so on the vendor's storage. The differences that fall out of that one choice are worth listing.

1. Retention is a setting, not a Facebook setting. Asyntai exposes configurable chat retention with auto-delete. The Facebook plugin's retention is whatever Facebook decides.
2. Access control is yours. Asyntai supports per-tab permissions and per-website scoping, so an agent on one site cannot see chats from another. The Facebook plugin does not have that distinction because the chats all live against the Page.
3. Audit log is yours. Asyntai's audit log is append-only. Entries cannot be edited or deleted from the UI. The Facebook Page inbox does not give the same guarantee.
4. PII redaction is yours. Asyntai's configurable PII redaction is a separate output classifier, not a prompt rule. The Facebook plugin relies on Facebook's own policies.
5. Single sign-on is yours. Asyntai supports OIDC single sign-on, including Entra and Google. A Facebook Page inbox is gated by Facebook logins.

These are not free differentiators. They are features you turn on, on a paid tier. They are also features you cannot get from a Facebook plugin, because the chat is not on your infrastructure.

How a small site chooses

Three profiles cover most of the people reading this.

1. A solo founder who takes the occasional question. The Facebook plugin is fine until it is not. A move to tawk.to's free tier removes the dependency on Facebook at no cost. The thread lives on tawk.to's servers, not Facebook's.
2. A small team that wants live chat with two or three agents. Crisp Mini at $45 a month covers four seats and is the simplest paid answer. LiveChat Starter at $19 per agent on annual billing is cheaper per seat if the team is small and the volume is steady.
3. A team that wants AI to answer the first 80% of questions and a human for the rest. Asyntai's Starter at $39 a month covers 2,500 messages, which is roughly 2,000 single-turn chats or several hundred multi-turn conversations. The transcript is on your account, scoped to your websites, and exportable from the dashboard.

None of these three is wrong. The point of the page is that the choice is between tools that own the conversation and the Facebook plugin, which does not.

How a busy site chooses

Busy is the harder question. The Facebook plugin does not bill you for being busy. A hosted widget does, and the bill shape depends on the meter.

A per-agent meter (LiveChat) makes the bill about your team. A per-conversation meter (Tidio) makes the bill about your visitors. A per-message meter (Asyntai) makes the bill about how much the AI actually says. A per-workspace meter (Crisp) makes the bill about how many seats you want.

The mistake most sites make is picking the meter that looks smallest on the entry tier and ignoring how it scales. A team that does 2,999 conversations a month pays $300 on Tidio Premium, which is conversation-metered and predictable; past that band the published ladder steps up again. The same team on LiveChat Starter at five agents is $95 a month on annual billing, which is per-agent and predictable in a different way.

Asyntai's Standard at $139 a month covers 15,000 messages, which is a higher ceiling than either, and the bill is on the AI's answer volume, not the visitor's question volume. A visitor who opens a chat and gets a single answer uses one message. A visitor who opens a chat and asks four follow-ups uses five.

What changes the answer

Five variables move the decision. None of them are in the published prices, because they are inputs you supply.

1. Visitor volume. A site that takes a hundred visitors a day is on a different tier than a site that takes a hundred visitors a minute. The Facebook plugin handles both. A hosted widget on a free tier handles the first and not the second.
2. Team size. A team of one fits a per-workspace meter. A team of ten fits a per-agent meter if volume is high, and a per-conversation meter if volume is spiky.
3. Question shape. A site whose questions are mostly answerable from existing content fits an AI meter (Asyntai). A site whose questions need a human every time fits a human meter (LiveChat, Crisp).
4. Opening hours. A site that only needs to answer in office hours can use any of the four. A site that needs to answer at 11pm needs a meter that does not bill per human shift.
5. Compliance posture. A site that needs SSO, audit logs, retention controls and PII redaction needs features that the Facebook plugin does not have, because the transcript is not on your infrastructure.

The mistake most people make

The mistake is to read "free" on the Facebook plugin and stop reading. The plugin is free. The cost is the loss of the transcript, the loss of the visitor's identity on your site, and the loss of any audit trail your team can read back.

The second mistake is to read "free" on tawk.to and assume the comparison is over. tawk.to's free tier is genuinely free, and the $19 is a branding fee. If the badge does not bother you, the question is answered.

The third mistake is to read a per-agent price and a per-conversation price as if they were the same number. They are not. Per agent, the bill is seats times dollars. Per conversation, the bill is chats times dollars. Per message, the bill is answers times dollars. Reading the smallest number on each row and treating it as the comparison is how teams overpay.

What to check on your own site

Three numbers answer the question without contacting a vendor.

1. How many chats did the Facebook plugin take last month? That number is the volume you need the new tool to absorb.
2. How many of those needed a real person to resolve? That number is the volume the AI cannot handle, and the minimum human capacity you need.
3. How many of those arrived outside office hours? That number is the volume the Facebook plugin did not handle, and the size of the answer the new tool has to give at 11pm.

Two of those three, on most small sites, are zero or near zero. That is the gap an AI live chat agent closes. The remaining number is the size of the human queue you still need, and the smallest per-agent tier that covers it.

Where to start

If the Facebook plugin is the only chat you have, the cheapest move is to do nothing until you have a reason. The plugin still works.

If the plugin's location is the problem, tawk.to's free tier is the lowest-friction answer. It is genuinely free, it removes the dependency on Facebook, and the thread is on your workspace. The trade-off is a badge in the widget.

If the badge is the problem, tawk.to at $19 a month removes it.

If you want AI to answer the first 80% and a human to handle the rest, Asyntai Starter at $39 a month covers 2,500 messages, and the transcript is on your account with per-website scoping and configurable retention. Above that, Standard at $139 a month covers 15,000 messages and Pro at $449 a month covers 50,000. Enterprise is a signed order form with a negotiated message allowance and is not sold self-serve.

If you want a small human team on a paid chat, Crisp Mini at $45 a month covers four seats and is the simplest paid answer. LiveChat Starter at $19 per agent per month on annual billing is cheaper per seat for very small teams.

If conversation volume is the main thing, Tidio's per-conversation meter makes the bill match the load. Free for 50 conversations, $24.17 for 100, $49.17 from 250, $300 for 2,999, $749 above that.

Sources