Why do teams move off self-hosted Chatwoot?
Almost nobody leaves because the software stopped working. They leave because of everything wrapped around it. Someone has to own the machine it runs on, apply updates, keep the database backed up, watch disk space before it fills, renew certificates, and work out why a webhook quietly stopped firing on the morning the inbox is busiest. None of that effort appears on an invoice, which is exactly why it gets underestimated at the moment you choose to self-host.
The second reason is staffing. Self-hosting feels fine while the person who set it up is still around and still has spare hours. When that person shifts to another project or leaves, support sits on a system nobody remaining wants to touch. A hosted product turns an internal dependency into a vendor relationship, and for a small team that trade is usually worth making.
The third reason is pace. The assistant layer inside support tools is changing fast, and on your own instance you own the upgrade path for every part of it: the model, the retrieval, the integrations. Hosted vendors ship those changes to you instead.
Is Chatwoot cloud the easiest answer?
For many teams, yes, and it deserves to be checked before anything else. The hosted version keeps the product your agents already know, with the same inbox model, the same labels, and the same conversation flow. No retraining, no rebuilding of habits, no arguing about where macros live. If your only complaint is the server, the shortest route is simply to stop running the server.
Check a few things before you treat it as the whole answer. Look at how agents are priced, because per-agent billing quietly shapes who you are willing to invite; a manager who looks in twice a week and a weekend helper both start to feel like line items rather than teammates. Look at which channels are included on the plan you would actually buy, not the plan at the top of the page. Look at how the AI features are packaged and metered, since that is where modern support pricing tends to hide its variable parts. If those answers suit you, you are finished and the rest of this post is optional reading.
Which hosted alternatives fit which kind of team?
There is no single best replacement, because Chatwoot users arrive from very different places. Match the tool to the shape of your support load.
Zendesk is the most mature ticketing system on this list, with deep workflow rules, routing, reporting and a very large app ecosystem. That depth is the reason to pick it and also the reason to think twice: it rewards teams that have someone to configure and maintain it, and it tends to be more product than a five-person team needs. If you are weighing it seriously, our Zendesk comparison walks through the trade-offs.
Intercom is strongest when support and product sit close together, with a polished messenger, in-app messaging and a capable AI agent. Read its pricing carefully, because seat charges and usage-based charges can both appear in the same bill depending on the plan you choose.
Freshdesk is a sensible landing spot for email-led teams who want familiar ticketing without a long implementation. It covers a lot of ground at its mid tiers and will feel recognisable to anyone who has run a shared queue before.
Gorgias is purpose-built for ecommerce. If most of your conversations are order status, returns and exchanges, having order context sitting next to the message is genuinely useful. Outside retail, that advantage mostly disappears.
Help Scout suits small teams who care about the quality of written replies. It is a clean shared inbox with docs attached, light on heavy automation by design, and pleasant for people who dislike feeling like they are operating a machine.
Tidio is widget-first and chat-led, a reasonable fit for small ecommerce and service sites that want a bot on the site quickly rather than a full ticketing backbone.
Corebee takes the opposite approach to pricing. It is one flat $99 per month with unlimited seats and unlimited AI conversations, so neither headcount nor volume changes the bill, and every channel is on the same plan. You can see the whole thing on the pricing page.
How do the pricing models actually differ?
Ignore headline figures for a moment and look at the shape of each model, because the shape is what decides your bill a year from now.
Per-seat pricing charges for people. It is predictable, and it quietly discourages the thing you want: putting more eyes on customer conversations. The product lead who would spot a recurring bug, the founder who wants to read tickets on Sunday, the seasonal hire you need for six weeks all become purchase decisions rather than obvious choices.
Per-resolution or per-conversation pricing charges for outcomes. It sounds fair and it reads well on a website, but it inverts your incentives: the better the AI performs, the more you owe, and a good month for deflection is a bad month for the bill. It also makes forecasting hard right when you most want a stable number, because a product launch or an outage spikes volume without warning.
Flat pricing charges for access. The bill stops being something you model and becomes something you forget about, which is the main reason a flat-rate inbox appeals to owners who do not want another variable cost to watch. The thing to verify is what the flat price excludes, since some vendors use a flat base and then meter the parts you actually plan to use.
What should you check about the AI before switching?
Every vendor on this list now advertises AI, so the advertising tells you nothing. Ask narrower questions instead.
Ask what the assistant reads. An assistant grounded in your website and your own knowledge base answers about your refund window and your delivery areas, rather than producing confident general prose that sounds right and is wrong. Thin source material is the most common reason an AI rollout disappoints.
Ask what happens when it does not know. The handoff matters more than the answers. The assistant should pass the conversation to a person when the customer asks for one and when its own confidence is low, carrying the full history so nobody has to repeat themselves.
Ask what the AI is allowed to do. Answering is table stakes; booking a meeting into Cal.com or Calendly and writing the contact back to HubSpot, Salesforce, Pipedrive, monday or Zoho is where it starts saving real work. Ask about voice too, and about disclosure: a receptionist that answers the phone should tell callers plainly that it is an AI.
Finally, ask what counts as billable. On usage-based plans, the definition of a resolution is the single most important sentence in the contract.
How many channels do you really need on day one?
Fewer than the feature grid suggests, but more than you have today. Start with the places customers already reach you: the website widget and the support mailbox for nearly everyone, plus the phone if calls go unanswered during busy hours. Add social and messaging channels where your customers actually live, which for many businesses means WhatsApp long before it means anything else.
What matters is whether the channels share one queue. Two inboxes and a separate phone log is not an omnichannel setup; it is three places to miss a message. Check whether channels arrive in the same thread view with the same history, and whether adding a new one later means a new plan, a new charge or a new integration project. A platform where every channel is already included removes a future negotiation you would rather not have.
What does moving off a self-hosted instance involve?
Less than people fear, mostly because the hard part is not the data. Export your conversation history for the record, but accept that you will rarely read it; what you genuinely need to carry across are canned replies, help articles, routing rules and the automations you rely on daily.
A practical sequence: write down the handful of workflows that must survive, then stand the new tool up alongside the old one with the widget pointed at a staging page. Load the knowledge base first, since the assistant is only as good as that content. Point one low-volume channel at the new system for a few days, read the transcripts closely, correct the sources where answers drift, then move the busier channels. Keep the old instance running read-only until you stop reaching for it, and only then switch it off.
Setup time varies with how much content you have ready. Standing up a hosted tool and connecting a first channel takes about ten minutes of real work on a product designed for it, with no developer involved; writing the knowledge base properly is the part that deserves an afternoon.
How do you choose without running a long evaluation?
Run a short, concrete test rather than a feature comparison. Take the twenty questions your team answers most often, including the awkward ones about refunds, delivery and edge-case policies, and ask each shortlisted tool to handle them from your own content. Read the replies as a customer would. Then invite a colleague who will not be a daily user and see whether the pricing model makes that invitation feel free or expensive.
Pick Chatwoot cloud if the product is right and only the hosting was wrong. Pick Zendesk or Freshdesk if process depth and ticketing structure are what you are missing. Pick Gorgias for retail, Help Scout for a small writing-led team, Intercom if product and support are the same motion, Tidio for a fast widget. Pick Corebee if the thing you want to stop thinking about is the bill, and a flat price with unlimited seats and unlimited AI conversations across every channel matches how you work.