Short answer: Tawk.to is well known as free live chat, and free is a genuinely compelling offer. The catch is not hidden fees, it is staffing. Live chat only converts when a human is sitting there ready to reply. If nobody is watching your chat at 8pm on a Sunday, an AI widget that answers and captures the lead does more for you than a free chat box nobody answers.
The real cost of live chat is attention
Every live-chat tool, free or paid, makes the same assumption: someone is available. When that holds, live chat is excellent. A real person typing beats any AI for nuance, empathy, and closing.
When it does not hold, a live-chat widget becomes a liability. A visitor opens it, types a question, waits, gets nothing, and leaves with a worse impression than if the widget had never been there. You have advertised availability and then not been available.
For a small service business, availability is exactly the thing you do not have. You are on a job, with a client, driving, or asleep. That is not a scheduling failure, it is the nature of the work.
So the honest question is not free versus paid. It is whether the tool needs your attention to work. For Tawk.to's current offering and what is included at each tier, check their site.
Who genuinely does well with free live chat
Pick free live chat if:
- Someone is reliably at a desk during your business hours and will actually watch the widget.
- Your enquiries cluster into those hours rather than evenings and weekends.
- You want humans handling every conversation and value that over speed of coverage.
- Budget is genuinely zero and staffing is genuinely available. That combination is the sweet spot.
- Your questions need judgement more than information, where a scripted or trained answer would be wrong anyway.
If that is you, free live chat is hard to argue with. Use it and spend the money elsewhere.
It is worth saying plainly: for free human live chat with your own staff, tawk.to is a strong product and nothing here is an argument that it is not. A tool that costs nothing and lets your receptionist talk to visitors in real time is a good deal by any measure. The question is only whether that receptionist exists and is watching.
The two-week decay
Here is the uncomfortable pattern, and almost everyone who has installed free live chat recognises it.
Week one, you watch the dashboard constantly. It is new, the notification sound is exciting, and you answer within seconds. Week two, you still catch most of them. By week four the notifications are muted because they fired during a job. By month three nobody has opened the dashboard. Six months later the widget is still on the site, still inviting people to chat, and messages go unanswered for hours or days.
The widget has not changed. Your attention has. And the tool is now actively costing enquiries, because a visitor who opens a chat box and gets silence forms a worse opinion of you than one who never saw a chat box at all.
Do the arithmetic yourself rather than trusting anyone else's. Look at your own chat transcripts from the last quarter. Count how many visitors opened a conversation. Count how many got a reply within five minutes. Count how many got a reply at all. Multiply the unanswered ones by whatever a job is worth to you. That number is what free is costing.
What happens when nobody is watching
The alternative is a widget that does not need you:
- It answers. Trained by crawling your website, so it knows your services, hours, coverage, and how you price. At any hour.
- It captures. Name, contact, and the nature of the enquiry, conversationally or through an in-chat form with your own fields.
- It tells you. Email, webhook, or Telegram notification the moment a lead comes in.
- It tracks. A leads inbox with New, Contacted, Qualified, Won or Lost, so follow-up is a process rather than a memory test.
- It steps aside when it should. Human handoff for conversations that need a person, and guardrails so it stays on your business.
That is Clerkzo. Not free forever, though there is a fourteen-day free trial, and paid starts at nineteen dollars a month flat with no per-conversation billing. The relevant comparison is not against zero, it is against the enquiries that currently vanish.
Be clear about the honest limit on our side too. Clerkzo is not a live-chat product. There is no agent console where your staff sit and type to visitors in real time. If what you want is your own people having live conversations, a live-chat tool is the right category and you should use one. What Clerkzo does is answer from your content and hand off to a human by capturing the enquiry and alerting you, which is a different shape of the same job.
What it looks like when it works
A useful way to picture the difference is to read the same conversation twice.
Unattended live chat, Sunday evening:
Visitor: Hi, do you do emergency callouts at the weekend?
No reply. Visitor closes the tab after ninety seconds and messages the next company on the results page.
Automated widget, same evening:
Visitor: Hi, do you do emergency callouts at the weekend?
Widget: Yes, we cover weekend emergencies across our service area. Weekend callouts are charged at our out-of-hours rate, and we confirm the price before any work starts. Do you want someone to call you back tonight?
Visitor: Yes please.
Widget: Great. What is the best number, and can you tell me briefly what has happened?
The second conversation ends with a named enquiry and a phone number in your inbox. Nobody was awake for either one.
How to evaluate any alternative
Answer quality. For live chat, quality equals whoever is typing, which is you. For an AI widget, test it: ask three hard questions including one your site does not cover, and check it admits the gap rather than inventing. Clerkzo gives you a playground to try this before going live.
Setup time. Both categories install fast. The difference is ongoing time. Live chat costs attention forever, an AI widget costs about ten minutes up front and a little correction afterwards. Done-for-you setup is available if you want zero.
Lead capture. Live chat captures whatever the agent remembers to ask. Automated capture is consistent, structured, and always delivered. Consistency matters more than people expect.
Billing model. Free is unbeatable on price, so be clear-eyed about what you are trading. On the paid side, watch for per-conversation charges. Clerkzo is flat, details on pricing.
Design control. Free tools sometimes brand the widget. Check whether that matters to you. Clerkzo has Widget Studio theming and Shadow DOM isolation so it looks like part of your site.
Coverage of the hours that matter. Look at when your enquiries actually arrive rather than when you assume they do. If a third of them land outside working hours, that is the part of the week the tool has to cover on its own.
A short checklist before you switch anything
Work through this honestly and the decision usually makes itself.
- Pull the last ninety days of chat transcripts. How many went unanswered?
- Note the times enquiries arrive. Are they inside your staffed hours?
- Ask who, by name, is responsible for watching the widget. If there is no name, there is no live chat.
- Write down the ten questions you get asked most. Are the answers already on your website?
- If they are not, fix that first. Any automated tool answers from your content, so a thin website produces thin answers.
- Test a trained widget in a playground with those ten questions before you put anything on your live site.
- Decide what a captured enquiry is worth to you, and compare that with a flat monthly fee rather than with zero.
Common mistakes when moving off free live chat
Running both in the same corner. Two widgets confuse visitors and split conversations. Pick one, remove the other properly rather than just hiding it.
Expecting the AI to close like your best salesperson. It answers and captures. The closing still happens on your call back, which is why speed matters more than polish.
Skipping the content work. Installing a trained widget on a website that does not state your prices, coverage, or hours produces a widget that cannot state your prices, coverage, or hours.
Turning off notifications. The whole point is that the enquiry reaches a human fast. If the alert goes to an inbox nobody reads, you have rebuilt the problem you were fixing.
Never reading the transcripts. The first fortnight of real conversations is the most useful content audit you will ever get. Read them.
Choose them if, choose Clerkzo if
Choose free live chat if someone will genuinely watch it during hours that match your enquiry pattern, you want humans on every conversation, and budget is the binding constraint.
Choose Clerkzo if nobody is reliably free to type, enquiries arrive outside working hours, and you would rather every conversation end with an answer and a captured lead than an unanswered message. Start with how it works, or read do I need a chatbot for my business if you are undecided.
Perfectly reasonable middle path: use Clerkzo as the always-on layer and let human handoff bring a person in when someone is around. That gets coverage without pretending you are always available. If the after-hours gap is the specific problem you are solving, the after-hours answering page covers that setup directly.
Frequently asked questions
Is an unanswered chat really worse than no chat?
For a first-time visitor, generally yes. It reads as neglect. If you cannot staff it, either automate it or turn it off.
Can I set the widget to only show during business hours?
You can, but consider whether that is backwards. After-hours is exactly when an automated widget earns its keep, because that is when nothing else is answering.
Will customers know they are talking to AI?
They should. Being upfront costs nothing and builds trust, and people are largely fine with it as long as the answers are useful and a human is reachable when needed.
What if the AI gets something wrong?
Read the transcripts in week one and use the corrections loop. Fixing a wrong answer once makes it stay fixed, which is a different maintenance model from retraining a person.
Can I keep tawk.to for daytime and add an AI widget for nights?
Technically you can run different tools at different times, but in practice two widgets on one site is a maintenance headache and a confusing experience. Most businesses are better off with one always-on widget plus human handoff, so a person joins when a person is available.
Does free live chat hurt my site speed?
Any third-party script has a cost. Judge it by measuring your own pages before and after rather than by vendor claims, and prefer widgets that load asynchronously so they never block your content rendering.