Eight Channels, One Inbox: A Monday-Morning Triage Test of WhatsMax

It is 8:47 on a Monday and the messages have already won. A customer asked about a delivery on WhatsApp at 7:12. Another asked the same question on Instagram at 7:40, because the first one had not been answered yet. A third wrote on Messenger, a fourth emailed, and somewhere in there a Telegram message about a refund is sitting unread because nobody on the team opens Telegram before coffee. Five apps, one question each, and no single place where anyone can see all of them at once. That is the morning WhatsMax is built for — or at least, that is the morning it claims to fix.

WhatsMax is a self-hosted script that puts eight messaging channels into one shared team inbox: WhatsApp, Messenger, Instagram, Telegram, WeChat, SMS, email, and a website chat widget. It costs $59 once on CodeCanyon, and the vendor is blunt about what that buys: the full source code, no monthly platform fee from them, no user limit in the script, and every integration included. The catch, which the listing is equally blunt about, is that you host it yourself on a VPS, and everyone else in the chain — Meta, your AI provider, your SMS provider, your payment gateway — bills you separately.

This post is not a feature list. It is a triage test. I opened the public demo the vendor publishes, took one message from each of the eight channels, and worked them the way a small business would on a Monday morning: sort first, answer second, and keep a note of everything that slows you down. The demo runs on seeded data, so the conversations are the vendor's own examples rather than live customers, and I will say so wherever it matters. What I am testing is the shape of the work — whether one inbox genuinely makes the pile smaller, or just moves it.

The test setup: one demo, eight strangers, four piles

The demo is public and the vendor prints the login on the listing itself: client@spagreen.net with the password 12345678, at whatsmax-demo.spagreen.net. No signup, no card, no sales call. That is already a point in its favour. Most messaging platforms hide the inbox behind a trial form and a calendar invite; here you can be inside the product in under a minute and see the thing you are actually buying.

Before touching a single message, I set the rules for the test. Every incoming message goes into one of four piles. Pile one: a human has to answer this now — money, anger, or a deadline. Pile two: the bot can answer this, because the answer already exists somewhere in the business. Pile three: waiting on us, not on the customer — an order that has not shipped, a refund that needs a decision. Pile four: noise, duplicates, and the customer who messaged on three channels about the same thing. The test is simple. If the inbox makes the sorting faster than eight apps would, it earns its keep. If it only makes the pile prettier, it does not.

I also decided what the test would not do. It would not send a real broadcast, because a demo is not the place to test a 50,000-contact send, and the listing's own broadcast claims are covered properly in the pre-send opt-in audit we published earlier. It would not test the SaaS billing side, because selling subscriptions to tenants is a different job from answering customers, and conflating the two is how inbox reviews turn into business-plan reviews. The question here is narrow on purpose: does one inbox make Monday morning easier for the people who answer?

One conversation, eight channels

The first thing the inbox does is the thing that sounds trivial and is not: it shows you everyone at once. WhatsApp, Messenger, Instagram, Telegram, WeChat, SMS, email, and the website widget all land in the same list, in the same order, with the channel marked on each conversation. You stop doing the morning tour — WhatsApp, then Instagram, then the email tab you forgot about — and start with a list.

The second thing it does is keep the customer in one piece. Open a conversation and the side panel shows the contact's attributes, tags, marketing consent, orders, and full cross-channel history. That is the feature that matters most in practice, because the Monday-morning pile is rarely eight different people. It is usually three people, one of whom messaged twice. The Instagram complaint and the WhatsApp follow-up belong to the same person, and the inbox knows it. The listing says contacts are merged across channels, and the v1.5.0 changelog adds automatic contact merging and a manual merge action — the kind of small, unglamorous feature that only exists because someone in production kept answering the same customer twice.

Working through the eight channels one by one, the differences between them are honest differences, not marketing ones. WhatsApp is where the money questions live, and it is the only channel where the reply has to go through Meta's template rules. Messenger and Instagram are where the vague questions live — "is this in stock?", "do you ship to...?" — the kind a bot answers well. Telegram and WeChat are where the regulars live, and regulars expect a human. SMS is where the short, urgent things arrive, and email is where the long, angry things arrive. A shared inbox does not flatten those differences. It just stops you from missing the email because you were busy in WhatsApp.

There is a limit worth naming here. The WhatsApp QR channel — the one that lets you connect an ordinary WhatsApp account by scanning a code instead of going through Meta's official Cloud API — ships disabled by default, behind an acknowledgement screen. The listing says why, in plain language: it uses a protocol Meta does not sanction, so the number can be banned. If your whole plan is "connect my personal WhatsApp number and start messaging," that channel is a trap door, not a feature. The official Cloud API route needs your own Meta app and business verification, which is real paperwork and real waiting. Neither route is free of friction. The honest version of this product is: eight channels, but one of them comes with a warning label and another comes with a form.

Where the bot earns its keep, and where it hands over

Every shared inbox eventually faces the same question, and it is not whether the bot is clever. It is whether the bot knows when to stop. WhatsMax's answer, according to the listing, has three parts: a chatbot grounded in your own documents through retrieval, auto-replies in every channel, and a per-conversation handover to a human. The v1.5.0 changelog adds the detail that makes the other two believable: an automation loop cap that hands the conversation to a human instead of letting the bot circle forever.

That loop cap is the most important line in the whole changelog, because it is the line that protects you from your own automation. A bot that answers "is this in stock?" from a document is genuinely useful — it is a question with a factual answer, and a human answering it for the fortieth time is a human not answering the refund. A bot that keeps answering a customer who is getting angrier, because nobody told it to stop, is how a small business loses a customer it never met. The difference between the two is not intelligence. It is a rule that says: after this many automated turns, a person takes over. WhatsMax has that rule. Whether it is set sensibly out of the box is a configuration question, and configuration questions are where small teams quietly fail — but the rule existing at all is the difference between a tool and a liability.

The bot's knowledge comes from your own documents, chunked and embedded, with OpenAI, Claude, or Gemini behind it — you bring your own API key, and the listing tracks token and cost use per workspace. That is the right design for a small business and also the honest one: the AI is not free, it is just billed by someone else. A shop that uploads its returns policy, its delivery times, and its price list will get a bot that answers the questions those documents answer. A shop that uploads nothing will get a bot that guesses, and guessing is worse than silence. The work of making the bot good is the work of writing the documents down, which is work most small businesses have been avoiding for years. The bot does not remove that work. It finally punishes you for skipping it.

The automation builder sits next to the bot, and it is the part of the product a small team will either love or never open. Drag-and-drop canvas, triggers for inbound messages, keywords, new contacts, tags, webhooks, orders placed or cancelled, abandoned carts, lead stage changes — and a one-sentence AI flow generator that turns a description into a ready-made flow. My opinion, after reading the trigger list twice: the abandoned-cart and order triggers are the ones that pay for the product, because they turn a message the customer already expects into a reply that arrives without anyone typing. The keyword triggers are the ones that quietly rot, because nobody revisits them when the business changes its mind about what a keyword means. Automations need a gardener. Budget an hour a month for it, or do not be surprised when the garden grows weeds.

The assignment test: does the right human get it?

An inbox with one person in it is a filing cabinet. An inbox with four people in it is a workplace, and workplaces need assignment. WhatsMax gives you assignment, labels, internal notes, and open, snoozed, and resolved states, plus canned replies and WhatsApp template sending from the composer. On paper that is the complete kit. In practice the test is narrower: when a message lands at 8:47, does the right person see it before 9:00, and does anyone else answer it by accident?

Assignment plus unread counters plus realtime updates is a reasonable answer to that test. A message assigned to one person stops being everyone's problem and starts being one person's job, and the internal note is where the first person tells the second person what they already promised. The snooze state is the underrated one — it is the difference between "I will deal with this after the delivery van arrives" and a message that sits in the open pile looking neglected until someone marks it resolved just to clear the screen. Resolved states only work if people trust them, and people only trust them if reopening is easy. That is a team habit, not a software feature, and no script can install it for you.

Canned replies deserve a word of caution. They are fast, and fast is what a Monday morning wants, but a canned reply that does not fit the question is worse than a slow reply. The composer lets you send templates and media, which is genuinely useful for the tenth "here is the tracking link" of the day. It is genuinely harmful for the first "I am sorry this arrived broken" of the day. My rule for a team of four: canned replies for facts, typed replies for feelings. Write that rule on the wall next to the monitor, because the software will not enforce it.

There is also the question of who is answering from where. The listing's roadmap promises a Flutter mobile app for agents, and the vendor already sells it separately as WhatsMax Agent — it is not included in the $59, and it needs WhatsMax 1.5 or later. On a Monday morning that matters less than it sounds, because the team is at desks. On a Saturday, when the owner is the only person answering and they are not at a desk, it matters a great deal. The web app is installable as a progressive web app with push notifications, which covers some of that gap. But if mobile answering is your whole workflow rather than your backup, check the add-on's price before you count the $59 as the whole cost.

What the inbox hides

Every product has a part it would rather you discovered later. With WhatsMax, the later discoveries are all in the listing if you read it properly, and they are all reasonable — but they are not small.

First, the hosting. This is a Node.js application, and it needs a VPS with Node 20 or newer, MySQL 8 or newer, Redis 6 or newer, an Nginx reverse proxy, and an SSL certificate. It needs a queue worker kept alive by PM2 and a scheduler tick from cron every minute. The listing says it plainly: shared cPanel or PHP hosting is not supported. If you are reading this on a site that runs on shared hosting — and you are — that line is aimed directly at you. The recommended server is 2 vCPU and 4 GB of RAM on Ubuntu. That is a real server, with a real monthly bill, and someone has to keep it updated. The $59 buys the code. The server is forever.

Second, the channels bill you. WhatsApp Cloud API conversations are priced by Meta. AI tokens are priced by OpenAI, Anthropic, or Google. SMS is priced per message by whichever of the fifteen providers you pick. Payment gateways take their transaction fees. The listing is honest about all of this — "WhatsMax itself never charges you again" — and that honesty is exactly why the total cost of running it is higher than $59. A small shop sending a few hundred messages a month will barely notice. A shop sending fifty thousand will notice very much, and should read Meta's conversation pricing before promising anyone a budget.

Third, the SaaS side is a second product wearing the same name. Plans, trials, coupons, tax rates, invoices, thirteen payment gateways, per-client white labelling, a super-admin panel with login-as-client, a landing-page builder, a CMS — all of it is in the box, and all of it is for someone who wants to sell messaging software to other businesses. That is a legitimate reason to buy this script, and it is the reason most buyers probably do buy it. But it is not the reason to buy it for your own shop's inbox, and the two reasons pull in opposite directions: the SaaS features are complexity you pay for in learning time even if you never switch them on. If you only want an inbox, you are buying a platform and using a corner of it. That is fine, as long as you know it.

The bill, in plain numbers

Here is the money, without adjectives. The script is $59 once, Regular License, on CodeCanyon. The Extended License costs more and buys you the same source code — the licences differ in how Envato lets you use the item, not in what you download. Support is six months, extendable at checkout, with a one-to-two business day response time on weekdays. Updates are lifetime for the purchased licence.

On top of that, budget a VPS — the listing recommends 2 vCPU and 4 GB of RAM — plus Meta's conversation pricing for WhatsApp, AI token usage for the chatbot, SMS per-message costs, and gateway fees on any payments you take. None of those numbers are in the listing, because none of them belong to the vendor. They belong to Meta, OpenAI, Twilio, and Stripe, and they change. Anyone who quotes you a single all-in monthly figure for running this is guessing. The honest budget is $59 once, a server every month, and usage that scales with how much you actually message.

Compared with renting a hosted inbox, the maths is simple and it cuts both ways. A hosted inbox charges you every month, forever, per seat or per conversation, and the bill grows with your team. WhatsMax charges you once and hands you the server bill instead. If you have the technical confidence — or a person — to run a Node.js app on a VPS, the one-time purchase wins within months. If you do not, the server will eat the savings in stress and in the first outage you have to fix yourself at 9pm. There is no wrong answer here, only a wrong assumption about which kind of buyer you are. The listing's own FAQ answers the shared-hosting question with a flat no, and I respect a vendor that would rather lose the sale than take the support ticket.

The verdict, after the pile is sorted

So: does one inbox make Monday morning easier? After working through the demo with eight channels and four piles, my answer is yes, but with a condition that the listing does not print on the box. The inbox earns its keep on the sorting, not on the answering. Seeing every channel in one list, seeing the same customer once instead of three times, and handing a conversation to the right person with a note attached — that is real time saved, every morning, and it compounds. The bot earns its keep on the factual questions and loses it the moment it has to handle a feeling. The assignment and snooze features earn their keep only if the team uses them, and no script can make a team do that.

The condition is the server. WhatsMax is not a product you buy; it is a product you adopt, and adoption means a VPS, a queue worker, a cron tick, and someone who knows what those words mean. For a small business that has that person — or is willing to become that person — $59 for a full omnichannel inbox with AI, automation, campaigns, and billing is a strikingly good deal, and the demo is public enough to verify every claim in this post before spending anything. For a business that does not have that person, the honest move is to rent a hosted inbox and pay monthly for someone else's server. The worst move is the middle one: buying the script, installing it on shared hosting that cannot run it, and discovering the FAQ's flat no after the money is spent.

If you want the day-in-the-life version of this product instead of the triage version, our earlier WhatsMax post follows one shop through a single day on one WhatsApp number. If your next problem is sending rather than answering, the broadcast opt-in audit walks through the list you should clean before you press send. And if the chatbot side is what interests you most, our audit of building a real week's AI tool stack is the closest thing on this site to a second opinion.

FAQ

Is WhatsMax a hosted service or a script I install?

It is a script you install. You buy the full source code once — $59 for the Regular License on CodeCanyon — and run it on your own VPS under your own brand. There is no monthly platform fee from the vendor and no user limit in the script. You pay Meta, your AI provider, your SMS provider, and your payment gateway directly for what you use.

Can I run WhatsMax on shared cPanel hosting?

No. It is a Node.js application that needs a VPS with Node.js 20 or newer, MySQL 8 or newer, Redis 6 or newer, an Nginx reverse proxy, and an SSL certificate. It also needs a queue worker kept alive by PM2 and a scheduler tick from cron every minute. The listing states plainly that shared cPanel or PHP hosting is not supported.

Which channels does the shared inbox cover?

Eight, plus a website chat widget: WhatsApp Cloud API, WhatsApp QR, Facebook Messenger, Instagram DM, Telegram, WeChat, SMS, and email. The WhatsApp QR channel ships disabled by default because it uses a protocol Meta does not sanction and the number can be banned. The official Cloud API route needs your own Meta app and business verification.

Does the AI chatbot answer customers on its own?

It can, and it is grounded in documents you upload, using OpenAI, Claude, or Gemini with your own API key. It also does auto-replies in every channel. The safeguard that matters is the handover: each conversation can be handed to a human, and the automation loop cap hands the conversation over rather than letting the bot circle. Upload good documents or the bot will guess, and guessing is worse than silence.

Is the mobile agent app included in the $59?

No. WhatsMax Agent, the Flutter app for iOS and Android, is sold separately, needs WhatsMax version 1.5 or later, and requires your own Apple and Google developer accounts to publish. The web app is installable as a progressive web app with push notifications, which covers day-to-day mobile answering for most small teams.

Can I sell WhatsMax as a SaaS to my own customers?

Yes — that is one of its two jobs. Plans, trials, coupons, tax rates, invoices, multi-currency billing, thirteen payment gateways, per-client white labelling, and a super-admin panel with login-as-client are all included in the source you download. The same licence buys the same code either way; the Regular and Extended licences differ only in how Envato lets you use the item.

Where can I try it before buying?

The vendor publishes a public demo at whatsmax-demo.spagreen.net with the login printed on the CodeCanyon listing — client@spagreen.net with the password 12345678 for the client app, and admin@spagreen.net with the same password for the super admin. Documentation is at whatsmax-docs.spagreen.net. Try the inbox in the demo before spending the $59, and try the server requirements on a spare VPS before promising anyone a launch date.